It is the most significant honor the Institute bestows on an individual and comes at a time Nadella described as a "golden age" of computer science.  

"I think what motivates all of us, as this community of folks who are associated with Georgia Tech, is not just the technology, because it's just merely a tool. But it's a powerful means to a more powerful end, which is about empowering every person and every organization on the planet to achieve more. The innovation you're driving here at Georgia Tech comes at one of the most consequential moments in the history of technology. As we enter this age of artificial intelligence, it's communities like this one that will help create the world we want to live in," Nadella said.  

With over 2,000 Georgia Tech alumni on its staff and a prominent presence in the city of Atlanta, Microsoft recently became the second company in history to surpass a $3 trillion market value. Accepting the degree, Nadella, who became the company's CEO in 2014, spoke of aligning values between Microsoft and Georgia Tech and looking forward to working together to create technologies and solutions for the world's most pressing challenges.  

"When you talk to Satya, he always leads with the impact that the company is having on people and organizations around the world," Georgia Tech President Ángel Cabrera said during the ceremony. "That sounds awfully familiar with our mission at Georgia Tech — to develop leaders who advance technology and improve the human condition. Our motto is progress and service. We define our success by the impact that we have in the lives of others. So, that explains why we're so excited to bring Satya into the family of Yellow Jackets." 

A fireside chat between Cabrera and Nadella followed the ceremony. Along with the transformative nature of AI and its ability to improve workflows and productivity in business, Nadella spoke of its potential to bring personalized instruction to students worldwide.  

He also participated in a roundtable discussion with faculty members about the implementation and impact of AI in higher education, robotics, cybersecurity computing, and other areas. Alex Orso, interim dean of the College of Computing, was among the participants and discussed ways to foster the exchange of talent between Georgia Tech and Microsoft. He also discussed how Georgia Tech's online computer science master's program, in collaboration with the Division of Lifetime Learning, can serve as a global platform to educate students before, during, and after college. 

Irfan Essa, senior associate dean in the College of Computing and co-director of Georgia Tech's AI Hub, added that Nadella and the faculty shared ideas for how industry and academia can collaborate to produce a strong workforce in the years to come.  

"We are a leader in this area and a part of the conversation," he said. "Industry leaders like Microsoft and places like Georgia Tech have to figure out a collaborative system to have more conversations about understanding the future workforce but also learning from companies about what kinds of things we should be providing from a broad standpoint educationally." 

Nadella then sat down with a group of students, all former Microsoft interns, to discuss their educational experience and what he called a "paradigm shift" across the industry, similar to the rise of the internet and cloud computing.  

Rynaa Grover will graduate with a master's degree in computer science in May and has accepted a position with Microsoft. "The research that goes on at Georgia Tech is very advanced and in line with the industry. It's incredible to be in this field at this point in time and to be able to contribute to such a big firm; it's really empowering," said Grover, who worked with Microsoft's machine learning platforms as an intern and with large language models in Assistant Professor Srijan Kumar’s lab. 

Yi Qin will also join Microsoft after graduating with a master's in human-computer interaction.  

"This was such a memorable experience for me,” she said. “Conversations like this make me feel like we are capable of doing a lot of great things. We should capture every opportunity that we have, have a growth mindset, and create whatever impact we want to make." 

A recent executive order issued by President Joe Biden seeks to establish "new standards for AI safety and security" while addressing consumer privacy concerns and promoting innovation. Georgia Tech experts have examined the key elements of the order and offer their thoughts on its scope and what comes next.  

A Precautionary Tale 

The order calls for the development of standards, tools, and tests to ensure the safe use of AI. From voice scams and phishing campaigns to larger-scale threats, the technology’s potential dangers have been widely documented. But Margaret Kosal, associate professor in the Ivan Allen College of Liberal Arts, says that additional context is often needed to dispel hysteria. 

"No one is going to be hooking up AI to launch nuclear weapons, but AI capabilities may enable targeting, or enable the command and control and the decision-making time to be compressed,” she said.  
 
The order will create an AI Safety and Security Board tasked with addressing critical threats. Companies developing foundation models that "pose a serious risk to national security, national economic security, or national public health and safety” will be required to notify the federal government when training the model and required to share the results of all red-team safety tests — a simulated cyberattack to test a system's defenses.  

Since the launch of ChatGPT in 2022, a CNBC report details a 1,267% rise in phishing emails. Srijan Kumar, assistant professor in the College of Computing, attributes the increase to the technology's availability and an inability to rein in "bad actors."  

He says these scams will only continue to get more sophisticated and personalized. They “can be created by knowing what you might be willing to fall prey to versus what I might fall prey to,” said Kumar, whose systems have influenced misinformation detection on sites like X (formerly Twitter) and Wikipedia. “AI is not going to autonomously do all of those bad things, but this order can ensure there are consequences for people who misuse it.”  

A Delicate Balance 

Building an AI platform requires large amounts of data regardless of its intended application. Two primary goals of the executive order are protecting privacy and advancing equity.  

To protect personal data, the order tasks Congress with evaluating how agencies collect and use commercially available information and address algorithmic discrimination.  

Acknowledging that everyone should be allowed to have their voice represented in the outputs of AI data sets, Deven Desai, associate professor in the Scheller College of Business, noted, "There are people who don't want to be part of data sets, which is their right, but this means their voices won't be reflected in the outputs.”   

The order also includes sections to address intellectual property concerns among inventors and creators, though legal challenges will likely set new precedents in the years ahead.  

When that time comes, Kosal says that defining “theft” in the context of AI becomes the true challenge and that, ultimately, money will play a significant role. "If you spit out a Harry Potter book and read it yourself, nobody will care. It's when you start selling it to make money, and you don't share proceeds with the original people, then it becomes an issue," she said.   

What Does AI-Generated Mean? 

The order instructs the Department of Commerce to develop guidelines for content authentication and watermarking to label AI-generated content. Desai questions what it means for something to be truly created by AI.  

An important distinction lies between using AI to assist a writer in organizing their thoughts and using the technology to generate content. He likens the trend to the music industry in the 1980s.  

"Synthesizers really changed people's ability to generate music and, for a while, people thought that was horrible. They can just program the music. They're not. I am still the human responsible for that music, or that article in this case, so what is the point of the label?" he asks. 

As AI assistance becomes commonplace in content creation, trusting the source of information is increasingly important. Recently, articles published on Sports Illustrated's website featured AI-generated content provided by a third-party company that had used a machine to write the content and create fake bylines. Sports Illustrated, which may not have known of the problem, ran the material without disclosure to readers. CEO Ross Levinsohn was ousted shortly after the story broke.  

“Perhaps if the third party had disclosed its use of AI software, SI would have been able to assess how much AI was used and then chosen not to run the material, or to run it with a disclaimer that AI helped write the material,” Desai said. "Of course, even if they label the content as AI-generated, a reader still won't know exactly how much of the content came from AI or a human.” 

AI and the Workforce 

As AI systems and models become more sophisticated, workers may become more concerned about being replaced. To counteract these concerns, the order calls for a study to examine AI’s potential impact on labor markets and investments in workforce training efforts.  

Kumar compares the rise of AI to similar technological innovations throughout history and sees it as an opportunity for workers and industries to adapt. "It's less a matter of AI replacing workers and more of reskilling people to use the new technology. It's no different from when assembly lines in the auto industry were created."  

Promoting Innovation and Competition 

The power to harness the full potential of AI has initiated a race to the top. Desai believes that part of the executive order providing resources to smaller developers can help level the playing field.   

"There is a possibility here for markets to open up. Current players using models that weren't built with transparency in mind might struggle, but maybe that's OK." 

The issue of reliability and transparency comes into focus for Desai, especially as it relates to government usage of AI. The order calls on agencies to "acquire specified AI products and services faster, more cheaply, and more effectively through more rapid and efficient contracting."  

When taxpayer dollars are at stake, government can’t afford to trust a technology it doesn’t fully understand — a topic Desai has explored elsewhere. "You can’t just say, ‘We don’t know how it works, but we trust it.’ That’s not going to work. So that’s where there may be a slowdown in the government’s ability to use private sector software if they can’t explain how the thing works and to show that it doesn’t have discriminatory issues.” 

What's Next 

Promoting and policing the safe use of AI cannot be done independently. Georgia Tech experts agree that participation on a global scale is necessary. To that end, the European Union will unveil its comprehensive EU AI Act, which includes a similar framework to the president's executive order.  

Due to the evolving nature of AI, the executive order or the EU's actions will not be all-encompassing. Law often lags behind technology, but Kosal points out that it's crucial to think beyond what currently exists when crafting policy.  

Experts also agree that AI cannot be regulated or governed through a single document and that this order is likely the first in a series of policymaking moves. Kosal sees tremendous opportunity with the innovation surrounding AI but hopes the growing fear of its rise does not usher in another AI winter, in which interest and research funding fade. 

Serving as a liaison between the community and responding officers, command center staff monitor calls, emails, and texts from Georgia Tech students, faculty, and staff. They also provide background information on suspects, research license plate numbers and warrants, and monitor the cameras on campus to provide video support for criminal investigations and to ensure officer safety during traffic stops.  

Shireka Graham, director of emergency communications, reminds all members of the Georgia Tech community not to hesitate to reach out to the operations center.  

"We want students, faculty, and staff to know that we are the heartbeat of the operation — they might not see us, but we are here for them,” she said. “We are the voice behind the call, and just like the police officers and other field responders, we're there with them.” 

Sometimes, she noted, people are reluctant to call because they don't think their situation is an emergency or that it's important. “But if they're in doubt, call us and let us get someone out there to evaluate the situation and get them the help they need," Graham said.  

"Our students are our top priority,” she said. “We know there is a lot of pressure” at Tech, especially during finals, “so we encourage our students to call us if they need someone to talk to. We can connect them with the Center for Mental Health Care and Resources, and if they call us, we will send a responder and remain on the line with them until help arrives." 

The operations center can be reached through a 911 call, the LiveSafe app (calls, texts, and audio and video messages), the center's administrative line for non-emergencies (404.894.2500), emergency elevator buttons, and 555 blue light emergency phones across campus. For those reaching out to the command center, Graham said it's important to share as much information as possible to help dispatchers provide the best response.  

Trained to Assist 

Before assuming their duties, dispatchers complete a rigorous 12-week training program, and ongoing training is required annually. The law enforcement arm of GTPD earned Commission on Accreditation for Law Enforcement Agencies (CALEA) accreditation in 2013, and Graham challenged her staff with earning the same accreditation — which they accomplished in 2022. The Operations Center is Georgia's only university or college communications center to receive it. 

“Being CALEA-accredited provides the best practices and standards to enhance our commitment to better serve and build trust within the Georgia Tech community,” Graham said. “The intensive process comprises 207 standards and 20-plus annual continued education training hours” for each dispatcher. “It's important to us to provide the highest level of customer service, process each call effectively and efficiently, and assure we're skilled to handle critical incidents regarding our students and the community we serve."   

Working alongside the dispatchers in the command center is a group of student assistants, and giving them an opportunity to see the inner workings of GTPD as video technicians is one of many outreach efforts to connect with the campus community.  

"Who knows the campus better than the students? They're the ones walking it every single day, so having them collaborate with us is very beneficial in meeting our common goal of ensuring safety for our campus," she said.   

In 2022, the team became the first university Public Safety Communications agency in the state to be recognized by the National Center for Missing and Exploited Children's Missing Kids & 9-1-1 Readiness Program. 

A new $2,000,000 grant from the National Science Foundation (NSF) will support the research. 

The research builds on previous collaboration between Rosa Arriaga, an associate professor in the College of Computing and Andrew Sherrill, an assistant professor in the Department of Psychiatry and Behavioral Sciences at Emory University, who worked together on a computational system for PTSD therapy. 

Arriaga and Christopher Wiese, an assistant professor in the School of Psychology will lead the Georgia Tech team, Saeed Abdullah, an assistant professor in the College of Information Sciences and Technology will lead the Penn State team, and Sherrill will serve as overall project lead and Emory team lead.

The grant, for “Understanding the Ethics, Development, Design, and Integration of Interactive Artificial Intelligence Teammates in Future Mental Health Work” will allocate $801,660 of support to the Georgia Tech team, supporting four years of research.

“The initial three years of our project are dedicated to understanding and defining what functionalities and characteristics make an AI system a 'teammate' rather than just a tool,” Wiese says. “This involves extensive research and interaction with mental health professionals to identify their specific needs and challenges. We aim to understand the nuances of their work, their decision-making processes, and the areas where AI can provide meaningful support.In the final year, we plan to implement a trial run of this AI teammate philosophy with mental health professionals.”

While the project focuses on mental health workers, the impacts of the project range far beyond. “AI is going to fundamentally change the nature of work and workers,” Arriaga says. “And, as such, there’s a significant need for research to develop best practices for integrating worker, work, and future technology.”

The team underscores that sectors like business, education, and customer service could easily apply this research. The ethics protocol the team will develop will also provide a critical framework for best practices. The team also hopes that their findings could inform policymakers and stakeholders making key decisions regarding AI. 

“The knowledge and strategies we develop have the potential to revolutionize how AI is integrated into the broader workforce,” Wiese adds. “We are not just exploring the intersection of human and synthetic intelligence in the mental health profession; we are laying the groundwork for a future where AI and humans collaborate effectively across all areas of work.”

Collaborative project

The project aims to develop an AI coworker called TEAMMAIT (short for “the Trustworthy, Explainable, and Adaptive Monitoring Machine for AI Team”). Rather than functioning as a tool, as many AI’s currently do, TEAMMAIT will act more as a human teammate would,  providing constructive feedback and helping mental healthcare workers develop and learn new skills.

“Unlike conventional AI tools that function as mere utilities, an AI teammate is designed to work collaboratively with humans, adapting to their needs and augmenting their capabilities,” Wiese explains. “Our approach is distinctively human-centric, prioritizing the needs and perspectives of mental health professionals… it’s important to recognize that this is a complex domain and interdisciplinary collaboration is necessary to create the most optimal outcomes when it comes to integrating AI into our lives.”

With both technical and human health aspects to the research, the project will leverage an interdisciplinary team of experts spanning clinical psychology, industrial-organizational psychology, human-computer interaction, and information science.

“We need to work closely together to make sure that the system, TEAMMAIT, is useful and usable,” adds Arriaga. “Chris (Wiese) and I are looking at two types of challenges: those associated with the organization, as Chris is an industrial organizational psychology expert — and those associated with the interface, as I am a computer scientist that specializes in human computer interaction.”

Long-term timeline

The project’s long-term timeline reflects the unique challenges that it faces.

“A key challenge is in the development and design of the AI tools themselves,” Wiese says. “They need to be user-friendly, adaptable, and efficient, enhancing the capabilities of mental health workers without adding undue complexity or stress. This involves continuous iteration and feedback from end-users to refine the AI tools, ensuring they meet the real-world needs of mental health professionals.”

The team plans to deploy TEAMMAIT in diverse settings in the fourth year of development, and incorporate data from these early users to create development guidelines for Worker-AI teammates in mental health work, and to create ethical guidelines for developing and using this type of system.

“This will be a crucial phase where we test the efficacy and integration of the AI in real-world scenarios,” Wiese says. “We will assess not just the functional aspects of the AI, such as how well it performs specific tasks, but also how it impacts the work environment, the well-being of the mental health workers, and ultimately, the quality of care provided to patients.”

Assessing the psychological impacts on workers, including how TEAMMAIT impacts their day-to-day work will be crucial in ensuring TEAMMAIT has a positive impact on healthcare worker’s skills and wellbeing.

“We’re interested in understanding how mental health clinicians interact with TEAMMAIT and the subsequent impact on their work,” Wiese adds. “How long does it take for clinicians to become comfortable and proficient with TEAMMAIT? How does their engagement with TEAMMAIT change over the year? Do they feel like they are more effective when using TEAMMAIT? We’re really excited to begin answering these questions.

A recent executive order issued by President Joe Biden seeks to establish "new standards for AI safety and security" while addressing consumer privacy concerns and promoting innovation. Georgia Tech experts have examined the key elements of the order and offer their thoughts on its scope and what comes next.  

A Precautionary Tale 

The order calls for the development of standards, tools, and tests to ensure the safe use of AI. From voice scams and phishing campaigns to larger-scale threats, the technology’s potential dangers have been widely documented. But Margaret Kosal, associate professor in the Ivan Allen College of Liberal Arts, says that additional context is often needed to dispel hysteria. 

"No one is going to be hooking up AI to launch nuclear weapons, but AI capabilities may enable targeting, or enable the command and control and the decision-making time to be compressed,” she said.  
 
The order will create an AI Safety and Security Board tasked with addressing critical threats. Companies developing foundation models that "pose a serious risk to national security, national economic security, or national public health and safety” will be required to notify the federal government when training the model and required to share the results of all red-team safety tests — a simulated cyberattack to test a system's defenses.  

Since the launch of ChatGPT in 2022, a CNBC report details a 1,267% rise in phishing emails. Srijan Kumar, assistant professor in the College of Computing, attributes the increase to the technology's availability and an inability to rein in "bad actors."  

He says these scams will only continue to get more sophisticated and personalized. They “can be created by knowing what you might be willing to fall prey to versus what I might fall prey to,” said Kumar, whose systems have influenced misinformation detection on sites like X (formerly Twitter) and Wikipedia. “AI is not going to autonomously do all of those bad things, but this order can ensure there are consequences for people who misuse it.”  

A Delicate Balance 

Building an AI platform requires large amounts of data regardless of its intended application. Two primary goals of the executive order are protecting privacy and advancing equity.  

To protect personal data, the order tasks Congress with evaluating how agencies collect and use commercially available information and address algorithmic discrimination.  

Acknowledging that everyone should be allowed to have their voice represented in the outputs of AI data sets, Deven Desai, associate professor in the Scheller College of Business, noted, "There are people who don't want to be part of data sets, which is their right, but this means their voices won't be reflected in the outputs.”   

The order also includes sections to address intellectual property concerns among inventors and creators, though legal challenges will likely set new precedents in the years ahead.  

When that time comes, Kosal says that defining “theft” in the context of AI becomes the true challenge and that, ultimately, money will play a significant role. "If you spit out a Harry Potter book and read it yourself, nobody will care. It's when you start selling it to make money, and you don't share proceeds with the original people, then it becomes an issue," she said.   

What Does AI-Generated Mean? 

The order instructs the Department of Commerce to develop guidelines for content authentication and watermarking to label AI-generated content. Desai questions what it means for something to be truly created by AI.  

An important distinction lies between using AI to assist a writer in organizing their thoughts and using the technology to generate content. He likens the trend to the music industry in the 1980s.  

"Synthesizers really changed people's ability to generate music and, for a while, people thought that was horrible. They can just program the music. They're not. I am still the human responsible for that music, or that article in this case, so what is the point of the label?" he asks. 

As AI assistance becomes commonplace in content creation, trusting the source of information is increasingly important. Recently, articles published on Sports Illustrated's website featured AI-generated content provided by a third-party company that had used a machine to write the content and create fake bylines. Sports Illustrated, which may not have known of the problem, ran the material without disclosure to readers. CEO Ross Levinsohn was ousted shortly after the story broke.  

“Perhaps if the third party had disclosed its use of AI software, SI would have been able to assess how much AI was used and then chosen not to run the material, or to run it with a disclaimer that AI helped write the material,” Desai said. "Of course, even if they label the content as AI-generated, a reader still won't know exactly how much of the content came from AI or a human.” 

AI and the Workforce 

As AI systems and models become more sophisticated, workers may become more concerned about being replaced. To counteract these concerns, the order calls for a study to examine AI’s potential impact on labor markets and investments in workforce training efforts.  

Kumar compares the rise of AI to similar technological innovations throughout history and sees it as an opportunity for workers and industries to adapt. "It's less a matter of AI replacing workers and more of reskilling people to use the new technology. It's no different from when assembly lines in the auto industry were created."  

Promoting Innovation and Competition 

The power to harness the full potential of AI has initiated a race to the top. Desai believes that part of the executive order providing resources to smaller developers can help level the playing field.   

"There is a possibility here for markets to open up. Current players using models that weren't built with transparency in mind might struggle, but maybe that's OK." 

The issue of reliability and transparency comes into focus for Desai, especially as it relates to government usage of AI. The order calls on agencies to "acquire specified AI products and services faster, more cheaply, and more effectively through more rapid and efficient contracting."  

When taxpayer dollars are at stake, government can’t afford to trust a technology it doesn’t fully understand — a topic Desai has explored elsewhere. "You can’t just say, ‘We don’t know how it works, but we trust it.’ That’s not going to work. So that’s where there may be a slowdown in the government’s ability to use private sector software if they can’t explain how the thing works and to show that it doesn’t have discriminatory issues.” 

What's Next 

Promoting and policing the safe use of AI cannot be done independently. Georgia Tech experts agree that participation on a global scale is necessary. To that end, the European Union will unveil its comprehensive EU AI Act, which includes a similar framework to the president's executive order.  

Due to the evolving nature of AI, the executive order or the EU's actions will not be all-encompassing. Law often lags behind technology, but Kosal points out that it's crucial to think beyond what currently exists when crafting policy.  

Experts also agree that AI cannot be regulated or governed through a single document and that this order is likely the first in a series of policymaking moves. Kosal sees tremendous opportunity with the innovation surrounding AI but hopes the growing fear of its rise does not usher in another AI winter, in which interest and research funding fade. 

Tutoring Without Borders is an independent platform helping to connect Ukrainian citizens in need of academic help with potential instructors or tutors.

“The opportunity to work with Tutoring Without Borders has been amazing. It’s been a great way to combine language and math, which are two completely different things, into something very meaningful,” Stephen said.

The first-year students grew up in Marietta, Georgia, and say that Georgia Tech was always their dream school — even though both of their parents went to UGA, which makes for a bit of lighthearted controversy in the family.

Christopher and Stephen both came to Tech as math majors but quickly gravitated toward computer science, finding in that discipline an appealing way to use math to solve tangible problems.

The computer science majors first became interested in Russian in high school. Both brothers were learning the violin, and their instructor was Russian. “He would always say little phrases in Russian, and one day our dad jokes and says, ‘Hey, why don’t you study Russian so you can be taught in Russian by your teacher?’” Christopher recalled.

What began as a joke inspired the brothers to enroll in Russian language courses in their first year of high school. They are now thinking about adding a Russian major to their resumes.

The Linders’ love of the Russian language has only deepened, and with the war in Ukraine, they felt compelled to help. “We felt a responsibility to help in some way because of our ability to communicate in Russian,” Stephen said.

Sharing their concerns with their Russian language teacher, they discovered  that math tutors are needed in Ukraine. Their teacher introduced them to Tutoring Without Borders, an independent platform connecting Ukrainian citizens with potential tutors.

Since June 2022, Christopher has been tutoring a young woman whose family was forced to move to Hungary. Stephen tutors two siblings who have been displaced to England. Regardless of where students are located, the brothers find time to assist them with their math — and even with their English.

Christopher and Stephen both note the challenges of teaching something technical in a different language, in addition to the emotional toll it can take working with students whose lives have been uprooted. Both continue to be passionate about helping and have no plans to stop anytime soon.

“I think just knowing that we can be there to provide some stability is very rewarding. And for me, personally, just getting to learn about my students and learn about their lives is something not many people are able to experience,” Christopher said.

For example, even an everyday chemical reaction, like toasting bread, can substantially change in speed and outcome — by increasing the heat, the speed of the reaction increases, toasting the bread faster. Adding another chemical ingredient — like buttering the bread before frying it — also changes the outcome of the reaction: the bread might brown and crisp rather than toast. The lesson? Certain chemical reactions can be accelerated or changed by adding or altering key variables, and understanding those factors is crucial when trying to create the desired reaction (like avoiding burnt toast!).

Chemists currently use quantum chemistry techniques to predict the rates and energies of chemical reactions, but the method is limited: predictions can usually only be made for up to a few hundred atoms. In order to scale the predictions to larger systems, and predict the environmental effects of reactions, a new framework needs to be developed.

Jesse McDaniel (School of Chemistry and Biochemistry) is creating that framework by leveraging computer modeling techniques. Now, a new NSF CAREER grant will help him do so. The National Science Foundation Faculty Early Career Development Award is a five-year grant designed to help promising researchers establish a foundation for a lifetime of leadership in their field. Known as CAREER awards, the grants are NSF’s most prestigious funding for untenured assistant professors. 

“I am excited about the CAREER research because we are really focusing on fundamental questions that are central to all of chemistry,” McDaniel says about the project.

Pioneering a new framework

“Chemical reactions are inherently quantum mechanical in nature,” McDaniel explains. “Electrons rearrange as chemical bonds are broken and formed.” While this type of quantum chemistry can allow scientists to predict the rates and energies of different reactions, these predictions are limited to only tens or hundreds of atoms. That’s where McDaniel’s team comes in. They’re developing modeling techniques based on quantum chemistry that could function over multiple scales, using computer models to scale the predictions. They hope this will help predict environmental effects on chemical reaction rates.

By developing modeling techniques that can be applied to reactions at multiple scales, McDaniel aims to expand scientist’s ability to predict and model chemical reactions, and how they interact with their environments. “Our goal is to understand the microscopic mechanisms and intermolecular interactions through which chemical reactions are accelerated within unique solvation environments such as microdroplets, thin films, and heterogenous interfaces,” McDaniel says. He hopes that it will allow for computational modeling of chemical reactions in much larger systems.  

Interdisciplinary research

As a theoretical and computational chemist, McDaniel’s chemistry experiments don’t take place in a typical chemistry lab — rather, they take place in a computer lab,  where Georgia Tech’s robust computer science and software development community functions as a key resource.

“We run computer simulations on high performance computing clusters,” McDaniel explains. “In this regard, we benefit from the HPC infrastructure at Georgia Tech, including the Partnership for an Advanced Computing Environment (PACE) team, as well as the computational resources provided in the new CODA building.” 

“Software is also a critical part of our research,” he continues. “My colleague Professor David Sherrill and his group are lead developers of the Psi4 quantum chemistry software, and this software comprises a core component of our multi-scale modeling efforts.”

In this respect, McDaniel is eager to to involve the next generation of chemists and computer scientists, showcasing the connection between these different fields. McDaniel’s team will partner with regional high school teachers, collaborating to integrate software and data science tools within the high school educational curriculum.

“One thing I like about this project,” McDaniel says, “is that all types of chemists — organic, inorganic, analytical, bio, physical, etc. — care about how chemical reactions happen, and how reactions are influenced by their surroundings.”

The Southeastern Cyber Cup is hosted by Georgia Tech’s Office of Information Technology in partnership with Deloitte. The virtual hacking event is open to cybersecurity and IT students and professionals and is held to generate enthusiasm and excitement around cybersecurity careers. 

As part of the annual competitors are challenged to find a "flag": a string of text. The flags for each challenge are submitted online to receive points. Challenge categories include network, web, crypto, miscellaneous, forensics, and reverse engineering.

Why Is the Southeastern Cyber Cup Important for GT/GTRI?

The Southeastern Cup and similar competitions are among the many ways that Georgia Tech and GTRI can showcase the skills of its researchers and aid in their professional development. The team’s Southeastern Cyber Cup win also indicates GTRI's role as a leader in the field of cybersecurity. 

“Historically, CTF (Capture the Flag) competitions are a practical way to sharpen the skills that any cybersecurity researcher/enthusiast may utilize during their career. If you’re interested in cybersecurity, CTFs are a great way to add new tools to your toolbox, as I often find myself picking up new skills during the course of such competitions,” Brown shared.

The Clockcycles team undoubtedly got the opportunity to sharpen their skills during the competition. Hsu shared that he and his team “stayed up for at least 20+ hours straight," participating in each event round. The time commitment and dedication certainly paid off in the end!

GT/GTRI's Impact on CTF Competitions

GTRI routinely has a group of researchers that participate in CTF competitions. In 2021, Petry and his team had an impressive win at the Hack-a-Sat 2 competition. In 2022, Petry and Hsu traveled to an east coast naval facility as part of a GTRI team that competed in person at an invitation-only event held by the US Navy, "Maritime Militia CTF." Their team was awarded a physical flag to bring back to GTRI, which they hung up as a trophy.

GTRI's dedication to these competitions hasn't gone unnoticed. At a CTF in 2022, GTRI received a letter of appreciation from the Naval Surface Warfare Center commending their performance. The Clockcyles' win at the Southeastern Cup is just one example of GTRI's impact as a research organization.

Meet the dedicated team members who brought home second place!

Justin Hsu

Justin Hsu is a Research Scientist in GTRI’S CIPHER (Cybersecurity, Information Protection, and Hardware Evaluation Research) Lab, Software Assurance Branch. Hsu's work includes looking at and working towards developing tools for software security testing and vulnerability analysis/assessments. He received a B.S. in Computer Information Systems from Shorter University, and an M.S. in Computer Science from Georgia Tech. Hsu has spent the majority of his professional career in software development. He previously worked at the ELSYS (Electronic Systems Laboratory) and shared that he moved to CIPHER after attending a seminar that rekindled his interest in cybersecurity. 

"I’ve been interested in cybersecurity since I was young, probably watching the movie ‘Hackers’ one too many times, and spent the majority of my career doing software development. But after hearing someone talk at a Friday Morning Seminar about their research work on malware, I was reminded of my interest in cybersecurity and wound up making the move from ELSYS to CIPHER," Hsu shared. 

This was the first year Hsu participated in the Southeastern Cyber Cup, but he has participated in CTFs with fellow CIPHER colleagues since 2021. To date, he's competed in about six different events, including ones sponsored by the U.S. Navy (HACKtheMACHINE, HACKtheMACHINE Unmanned) and the U.S. Air Force/Space Force (Hack-a-Sat, Hack-a-Sat 2, and Hack-a-Sat 3). 

Drew Petry

Drew Petry works as a Research Engineer in the Embedded System Vulnerability Division (ESVD) of CIPHER. Petry’s work focuses on the reverse engineering and security assessment of embedded systems and cyber EW (Electronic Warfare) techniques. He received a B.S. in Computer Engineering from Georgia Tech in 2010. In 2014, he also received his M.S. in Electrical and Computer Engineering from Georgia Tech.

Petry has spent the past fourteen years as a professional research engineer at GTRI, working in the embedded system security and vulnerability field. He shared that he’s always been drawn to embedded systems because he "enjoys interacting with low-level hardware and ‘bare-metal’ code.” Bare metal programming is the process of programming directly on the hardware without using an operating system or middleware.

Outside of the inaugural Southeastern Cyber Cup competition, Petry competes in capture-the-flag competitions yearly. The events he’s competed in while representing GTRI include the annual U.S. Air/Space Force Hack-a-sat CTFs and the U.S. Navy Hack the Machine cybersecurity competitions.

Garrett Brown

Garrett Brown is a Research Scientist in the Embedded Cyber Techniques (ECT) branch of the ESVD at CIPHER. He primarily works on vulnerability discovery and analysis of embedded systems. Brown received his B.S. in Computer Science from Georgia Tech. 

Brown shared that he found his passion in this field after participating in the VIP (Vertically Integrated Project) program while at Georgia Tech as an undergraduate student. During this program, he was a part of the Embedded Systems Cyber Security (ESCS) team, which gave him his "first taste of the work [he] would soon come to love."

"I believe cybersecurity practitioners can improve the lives of many around the world, and I'd like to be a part of whatever positive impact we can make," shared Brown when asked why he was passionate about his work.

While this was Brown's first time competing in the Southeastern Cyber Cup, he is not a stranger to competitions. He's previously competed in other CTFs as part of the CIPHER team for competitions such as the Hack-a-Sat and HACKtheMACHINE events.

When asked how he felt about their team's award, he shared, "I felt both relief and disappointment--relief that I could finally go to sleep and disappointment that we got second place instead of first!”

Congratulations Clockcycles team!

Writer: Madison McNair (madison.mcnair@gtri.gatech.edu)  
Photographer: Christopher Moore 
GTRI Communications  
Georgia Tech Research Institute  
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $940 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

The Georgia Tech Research Institute (GTRI) and Georgia Institute of Technology (Georgia Tech) are exploring how the powerful processing capabilities of quantum computers can expedite CFD’s resource-intensive simulations used in aircraft design, weather prediction, nuclear weapons testing and more.  

“Through a collaboration between GTRI and Georgia Tech, we are developing an application of quantum computing to solve proof-of-principle problems in computational fluid dynamics that could streamline efficiencies and reduce costs across numerous industries,” said Bryan Gard, a GTRI senior research scientist who is leading this project.

Quantum computing offers a new way of doing computations using the principles of quantum mechanics, a science that explores the behavior of tiny particles such as atoms and photons. Computers and software that are built on the theories of quantum mechanics can process a large amount of information simultaneously and much faster than classical computers. That is because unlike classical computers, which use bits that are either 0 or 1, quantum computers use quantum bits or qubits. 

Classical bits are similar to regular on/off switches, which can only exist in one state at a time. Qubits, meanwhile, can exist in multiple states at once thanks to a property in quantum mechanics known as superposition.  

Because CFD involves complex simulations of how fluids, such as air or water, move and interact with different surfaces, classical computers often struggle with the immense number of calculations needed for such detailed simulations. The ability for quantum computers to process information in parallel could significantly speed up these simulations and produce more accurate results. 

“Say you are examining how air flows over a plane wing and you want to identify the large- and small-scale dynamics of that interaction,” explained Gard. “This type of problem would be very hard for a classical computer to handle because it wouldn’t be able to examine those large- and small-scale aspects simultaneously.” 

The team has split its research into two parts. The parts that involve linear differential equations are solved on a quantum computer and the other, non-linear parts are handled conventionally on a classical machine. 

The reason for this division is that as the problem scales up on classical supercomputers, the communication between nodes becomes inefficient, creating a bottleneck. Even though quantum computers are not yet large-scale, they can handle certain parts of the problem without facing the same communication challenges, Gard explained. 

These principles could help organizations strategically allocate resources and avoid costs associated with manufacturing and testing potentially flawed designs. In the defense realm, an example of this can be seen with designing aircraft. 

Instead of the conventional methods of building and testing structures in a wind tunnel, quantum-enhanced CFD would allow engineers to analyze stresses, assess designs and predict performance more efficiently and cost effectively. This becomes particularly relevant at high speeds, where factors such as air flows and turbulence pose additional challenges for running accurate simulations. 

“It all comes down to money, as with everything else,” said Gard. “If you could save yourself a lot of time and money by running this simulation, which you couldn't do before, then it would allow you to allocate your resources more effectively.” 

For this project, GTRI is collaborating with Spencer Bryngelson, an assistant professor in the School of Computational Science and Engineering who has expertise in computational physics, numerical methods, fluid dynamics and high-performance computing. Zhixin Song, a graduate student at Georgia Tech who is researching quantum algorithms for CFD, has also contributed.   

“This project is particularly interesting because although it is challenging, it could have outsize performance gains if one can find the right tools for the job, meaning the right quantum algorithm to solve the right fluid dynamics problem,” Bryngelson said. “GTRI and Georgia Tech have already made progress in this area, and also work well together, so it has been a good experience.” 

The project has been supported by GTRI’s Independent Research and Development (IRAD) Program, winning an IRAD of the Year award in fiscal year 2023, and the Defense Advanced Research Projects Agency (DARPA). 

Writer: Anna Akins 
Photos: Christopher Moore 
Art Credit: Img2Go.com, Adobe 
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $940 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

Honolulu Highlights | ICML 2023
Students and faculty have been focused and energized in their efforts this week engaging with the international machine learning community at ICML. See some of those efforts, hear from students themselves in our video series, and read about their latest contributions in #AI.

Georgia Tech’s experts and larger research community are invested in a future where artificial intelligence (AI) solutions can benefit individuals and communities across our planet. Meet the machine learning maestros among Georgia Tech’s faculty at the International Conference on Machine Learning — July 23-29, 2023, in Honolulu — and learn about their work. The faculty in the main program are working with partners across many domains and industries to help invent powerful new ways for technology to benefit all our futures.

One of the experts in Honolulu is Wenjing Liao, an assistant professor in the School of Mathematics. In addition to machine learning, Liao's research interests include imaging, signal processing, and high dimensional data analysis.

Learn more about the Georgia Tech contingent at the ICML here. Read more about machine learning research at Georgia Tech here.

Improving these models is critical: while melting ice sheets and glaciers are top contributors to sea level rise, there are still large uncertainties in sea level projections at 2100 and beyond.

“Part of the problem is that the way that many models have been coded in the past has not been conducive to using these kinds of tools,” Robel, an assistant professor in the School of Earth and Atmospheric Sciences, explains. “It's just very labor-intensive to set up these data assimilation tools — it usually involves someone refactoring the code over several years.”

“Our goal is to provide a tool that anyone in the field can use very easily without a lot of labor at the front end,” Robel says. “This project is really focused around developing the computational tools to make it easier for people who use ice sheet models to incorporate or inform them with the widest possible range of measurements from the ground, aircraft and satellites.”

Now, a $780,000 NSF CAREER grant will help him to do so. 

The National Science Foundation Faculty Early Career Development Award is a five-year funding mechanism designed to help promising researchers establish a personal foundation for a lifetime of leadership in their field. Known as CAREER awards, the grants are NSF’s most prestigious funding for untenured assistant professors.

“Ultimately,” Robel says, “this project will empower more people in the community to use these models and to use these models together with the observations that they're taking.”
 

Ice sheets remember

“Largely, what models do right now is they look at one point in time, and they try their best — at that one point in time — to get the model to match some types of observations as closely as possible,” Robel explains. “From there, they let the computer model simulate what it thinks that ice sheet will do in the future.”

In doing so, the models often assume that the ice sheet starts in a state of balance, and that it is neither gaining nor losing ice at the start of the simulation. The problem with this approach is that ice sheets dynamically change, responding to past events — even ones that have happened centuries ago. “We know from models and from decades of theory that the natural response time scale of thick ice sheets is hundreds to thousands of years,” Robel adds.

By informing models with historical records, observations, and measurements, Robel hopes to improve their accuracy. “We have observations being made by satellites, aircraft, and field expeditions,” says Robel. “We also have historical accounts, and can go even further back in time by looking at geological observations or ice cores. These can tell us about the long history of ice sheets and how they've changed over hundreds or thousands of years.”

Robel’s team plans to use a set of techniques called data assimilation to adjust, or ‘nudge’, models. “These data assimilation techniques have been around for a really long time,” Robel explains. “For example, they’re critical to weather forecasting: every weather forecast that you see on your phone was ultimately the product of a weather model that used data assimilation to take many observations and apply them to a model simulation.”

“The next part of the project is going to be incorporating this data assimilation capability into a cloud-based computational ice sheet model,” Robel says. “We are planning to build an open source software package in Python that can use this sort of data assimilation method with any kind of ice sheet model.”

Robel hopes it will expand accessibility. “Currently, it's very labor-intensive to set up these data assimilation tools, and while groups have done it, it usually involves someone re-coding and refactoring the code over several years.”

Building software for accessibility

Robel’s team will then apply their software package to a widely used model, which now has an online, browser-based version. “The reason why that is particularly useful is because the place where this model is running is also one of the largest community repositories for data in our field,” Robel says.

Called Ghub, this relatively new repository is designed to be a community-wide place for sharing data on glaciers and ice sheets. “Since this is also a place where the model is living, by adding this capability to this cloud-based model, we'll be able to directly use the data that's already living in the same place that the model is,” Robel explains. 

Users won’t need to download data, or have a high-speed computer to access and use the data or model. Researchers collecting data will be able to upload their data to the repository, and immediately see the impact of their observations on future ice sheet melt simulations. Field researchers could use the model to optimize their long-term research plans by seeing where collecting new data might be most critical for refining predictions.

“We really think that it is critical for everyone who's doing modeling of ice sheets to be doing this transient data simulation to make sure that our simulations across the field are all doing the best possible job to reproduce and match observations,” Robel says. While in the past, the time and labor involved in setting up the tools has been a barrier, “developing this particular tool will allow us to bring transient data assimilation to essentially the whole field.”

Bringing Real Data to Georgia’s K-12 Classrooms

The broad applications and user-base expands beyond the scientific community, and Robel is already developing a K-12 curriculum on sea level rise, in partnership with Georgia Tech CEISMC Researcher Jayma Koval. “The students analyze data from real tide gauges and use them to learn about statistics, while also learning about sea level rise using real data,” he explains.

Because the curriculum matches with state standards, teachers can download the curriculum, which is available for free online in partnership with the Southeast Coastal Ocean Observing Regional Association (SECOORA), and incorporate it into their preexisting lesson plans. “We worked with SECOORA to pilot a middle school curriculum in Atlanta and Savannah, and one of the things that we saw was that there are a lot of teachers outside of middle school who are requesting and downloading the curriculum because they want to teach their students about sea level rise, in particular in coastal areas,” Robel adds.

In Georgia, there is a data science class that exists in many high schools that is part of the computer science standards for the state. “Now, we are partnering with a high school teacher to develop a second standards-aligned curriculum that is meant to be taught ideally in a data science class, computer class or statistics class,” Robel says. “It can be taught as a module within that class and it will be the more advanced version of the middle school sea level curriculum.”

The curriculum will guide students through using data analysis tools and coding in order to analyze real sea level data sets, while learning the science behind what causes variations and sea level, what causes sea level rise, and how to predict sea level changes. 

“That gets students to think about computational modeling and how computational modeling is an important part of their lives, whether it's to get a weather forecast or play a computer game,” Robel adds. “Our goal is to get students to imagine how all these things are combined, while thinking about the way that we project future sea level rise.”

The National Science Foundation Faculty Early Career Development Award is a five-year grant designed to help promising researchers establish a foundation for a lifetime of leadership in their field. Known as CAREER awards, the grants are NSF’s most prestigious funding for untenured assistant professors.

Bernshteyn, an assistant professor in the School of Mathematics, will focus on “Developing a unified theory of descriptive combinatorics and local algorithms” — connecting concepts and work being done in two previously separate mathematical and computer science fields. “Surprisingly,” Bernshteyn says, “it turns out that these two areas are closely related, and that ideas and results from one can often be applied in the other.” 

“This relationship is going to benefit both areas tremendously,” Bernshteyn says. “It significantly increases the number of tools we can use”

By pioneering this connection, Bernshteyn hopes to connect techniques that mathematicians use to study infinite structures (like dynamic, continuously evolving  structures found in nature), with the algorithms computer scientists use to model large – but still limited – interconnected networks and systems (like a network of computers or cell phones).

“The final goal, for certain types of problems,” he continues, “is to take all these questions about complicated infinite objects and translate them into questions about finite structures, which are much easier to work with and have applications in practical large-scale computing.”

Creating a unified theory

It all started with a paper Bernshteyn wrote in 2020, which showed that mathematics and computer science could be used in tandem to develop powerful problem-solving techniques. Since the fields used different terminology, however, it soon became clear that a “dictionary” or a unified theory would need to be created to help specialists communicate and collaborate. Now that dictionary is being built, bringing together two previously-distinct fields: distributed computing (a field of computer science), and descriptive set theory (a field of mathematics). 

Computer scientists use distributed computing to study so-called “distributed systems,” which model extremely large networks — like the Internet — that involve millions of interconnected machines that are operating independently (for example, blockchain, social networks, streaming services, and cloud computing systems).

“Crucially, these systems are decentralized,” Bernshteyn says. ”Although parts of the network can communicate with each other, each of them has limited information about the network’s overall structure and must make decisions based only on this limited information.” Distributed systems allow researchers to develop strategies — called distributed algorithms — that “enable solving difficult problems with as little knowledge of the structure of the entire network as possible,” he adds.

At first, distributed algorithms appear entirely unrelated to the other area Bernshteyn’s work brings together: descriptive set theory, an area of pure mathematics concerned with infinite sets defined by “simple” mathematical formulas. 

“Sets that do not have such simple definitions typically have properties that make them unsuitable for applications in other areas of mathematics. For example, they are often non-measurable – meaning that it is impossible, even in principle, to determine their length, area, or volume," Bernshteyn says.

Because undefinable sets are difficult to work with, descriptive set theory aims to understand which problems have “definable”— and therefore more widely applicable— solutions. Recently, a new subfield called descriptive combinatorics has emerged. “Descriptive combinatorics focuses specifically on problems inspired by the ways collections of discrete, individual objects can be organized,” Bernshteyn explains. “Although the field is quite young, it has already found a number of exciting applications in other areas of math.”

The key connection? Since the algorithms used by computer scientists in distributed computing are designed to perform well on extremely large networks, they can also be used by mathematicians interested in infinite problems.

Solving infinite problems

Infinite problems often occur in nature, and the field of descriptive combinatorics has been particularly successful in helping to understand dynamical systems: structures that evolve with time according to specified laws (such as the flow of water in a river or the movement of planets in the Solar System). “Most mathematicians work with continuous, infinite objects, and hence they may benefit from the insight contributed by descriptive set theory,” Bernshteyn adds.

However, while infinite problems are common, they are also notoriously difficult to solve. “In infinite problems, there is no software that can tell you if the problem is solvable or not. There are infinitely many things to try, so it is impossible to test all of them. But if we can make our problems finite, we can sometimes determine which ones can and cannot be solved efficiently,” Bernshteyn says. “We may be able to determine which combinatorial problems can be solved in the infinite setting and get an explicit solution.”

“It turns out that, with some work, it is possible to implement the algorithms used in distributed computing on infinite networks, providing definable solutions to various combinatorial problems,” Bernshteyn says. “Conversely, in certain limited settings it is possible to translate definable solutions to problems on infinite structures into efficient distributed algorithms — although this part of the story is yet to be fully understood.”

A new frontier

As a recently emerged field, descriptive combinatorics is rapidly evolving, putting Bernshteyn and his research on the cutting edge of discovery. “There’s this new communication between separate fields of math and computer science—this huge synergy right now—it’s incredibly exciting,” Bernshteyn says.

Introducing new researchers to descriptive combinatorics, especially graduate students, is another priority for Bernshteyn. His CAREER grant funds will be especially dedicated to training graduate students who might not have had prior exposure to descriptive set theory. Bernshteyn also aims to design a suite of materials ranging from textbooks, lecture notes, instructional videos, workshops, and courses to support students and scholars as they enter this new field.

“There’s so much knowledge that’s been acquired,” Bernshteyn says. “There’s work being done by people within computer science, set theory, and so on. But researchers in these fields speak different languages, so to say, and a lot of effort needs to go into creating a way for them to understand each other. Unifying these fields will ultimately allow us to understand them all much better than we did before. Right now we’re only starting to glimpse what’s possible.”

In a new paper published in eLife, researchers in the School of Biological Sciences and the School of Computer Science have shown that AF2Complex may be able to lend a hand.

Building on the models of DeepMind’s AlphaFold 2, a machine learning tool able to predict the detailed three-dimensional structures of individual proteins, AF2Complex — short for AlphaFold 2 Complex — is a deep learning tool designed to predict the physical interactions of multiple proteins. With these predictions, AF2Complex is able to calculate which proteins are likely to interact with each other to form functional complexes in unprecedented detail.

“We essentially conduct computational experiments that try to figure out the atomic details of supercomplexes (large interacting groups of proteins) important to biological functions,” explained Jeffrey Skolnick, Regents’ Professor and Mary and Maisie Gibson Chair in the School of Biological Sciences, and one of the corresponding authors of the study. With AF2Complex, which was developed last year by the same research team, it’s “like using a computational microscope powered by deep learning and supercomputing.”

In their latest study, the researchers used this ‘computational microscope’ to examine a complicated protein synthesis and transport pathway, hoping to clarify how proteins in the pathway interact to ultimately transport a newly synthesized protein from the interior to the outer membrane of the bacteria — and identify players that experiments might have missed. Insights into this pathway may identify new targets for antibiotic and therapeutic design while providing a foundation for using AF2Complex to computationally expedite this type of biology research as a whole.

Computing complexes

Created by London-based artificial intelligence lab DeepMind, AlphaFold 2 is a deep learning tool able to generate accurate predictions about the three-dimensional structure of single proteins using just their building blocks, amino acids. Taking things a step further, AF2Complex uses these structures to predict the likelihood that proteins are able to interact to form a functional complex, what aspects of each structure are the likely interaction sites, and even what protein complexes are likely to pair up to create even larger functional groups called supercomplexes.

“The successful development of AF2Complex earlier this year makes us believe that this approach has tremendous potential in identifying and characterizing the set of protein-protein interactions important to life,” shared Mu Gao, a senior research scientist at Georgia Tech. “To further convince the broad molecular biology community, we [had to] demonstrate it with a more convincing, high impact application.”

The researchers chose to apply AF2Complex to a pathway in Escherichia coli (E. coli), a model organism in life sciences research commonly used for experimental DNA manipulation and protein production due to its relative simplicity and fast growth. 

To demonstrate the tool’s power, the team examined the synthesis and transport of proteins that are essential for exchanging nutrients and responding to environmental stressors: outer membrane proteins, or OMPs for short. These proteins reside on the outermost membrane of gram-negative bacteria, a large family of bacteria characterized by the presence of inner and outer membranes, like E. coli. However, the proteins are created inside the cell and must be transported to their final destinations. 

“After more than two decades of experimental studies, researchers have identified some of the protein complexes of key players, but certainly not all of them,” Gao explained. AF2Complex “could enable us to discover some novel and interesting features of the OMP biogenesis pathway that were missed in previous experimental studies.”

New insights

Using the Summit supercomputer at the Oak Ridge National Laboratory, the team, which included computer science undergraduate Davi Nakajima An, put AF2Complex to the test. They compared a few proteins known to be important in the synthesis and transport of OMPs to roughly 1,500 other proteins — all of the known proteins in E. coli’s cell envelope — to see which pairs the tool computed as most likely to interact, and which of those pairs were likely to form supercomplexes. 

To determine if AF2Complex’s predictions were correct, the researchers compared the tool’s predictions to known experimental data. “Encouragingly,” said Skolnick, “among the top hits from computational screening, we found previously known interacting partners.” Even within those protein pairs known to interact, AF2Complex was able to highlight structural details of those interactions that explain data from previous experiments, lending additional confidence to the tool’s accuracy.

In addition to known interactions, AF2Complex predicted several unknown pairs. Digging further into these unexpected partners revealed details on what aspects of the pairs might interact to form larger groups of functional proteins, likely active configurations of complexes that have previously eluded experimentalists, and new potential mechanisms for how OMPs are synthesized and transported. 

“Since the outer membrane pathway is both vital and unique to gram-negative bacteria, the key proteins involved in this pathway could be novel targets for new antibiotics,” said Skolnick. “As such, our work that provides molecular insights about these new drug targets might be valuable to new therapeutic design.”

Beyond this pathway, the researchers are hopeful that AF2Complex could mean big things for biology research. 

“Unlike predicting structures of a single protein sequence, predicting the structural model of a supercomplex can be very complicated, especially when the components or stoichiometry of the complex is unknown,” Gao noted. “In this regard, AF2Complex could be a new computational tool for biologists to conduct trial experiments of different combinations of proteins,” potentially expediting and increasing the efficiency of this type of biology research as a whole.

AF2Complex is an open-source tool available to the public and can be downloaded here.

This work was supported in part by the DOE Office of Science, Office of Biological and Environmental Research (DOE DE-SC0021303) and the Division of General Medical Sciences of the National Institute Health (NIH R35GM118039). DOI: https://doi.org/10.7554

STEM@GTRI celebrated its 25th anniversary recently. STEM @GTRI is the Georgia Tech Research Institute's K-12 outreach program. STEM @GTRI strives to inspire, engage, and impact Georgia's students and educators through hands-on experiences, outreach, and professional learning.

STEM@GTRI customizes professional development experiences for educators, connects students and classrooms to Georgia Tech labs and researchers, and brings hands-on, fun, and relevant programming to STEM (science, technology, engineering, and math) educational outreach events across Georgia. STEM@GTRI leverages State of Georgia funding through grants and partnerships to bring additional STEM programming to K-12 students in Georgia. The program first received State of Georgia funding in 1998.

To commemorate this auspicious occasion, STEM@GTRI hosted a luncheon celebrating 25 years of K-12 STEM outreach at GTRI. During the program, an array of speakers reflected on the STEM @GTRI program over the past 25 years and its impact in Georgia and on the future of students.

STEM @GTRI’s First Champion: Claudia Huff

Claudia Huff, the retired GTRI Principal Research Associate who was the first Director of STEM @GTRI, spoke on its inspirational and aspirational early days. She noted that, in 1998, the U.S. was experiencing a rapid permeation of emphasis on STEM education, fueled by legislation such as the Telecommunications Act of 1996. However, while there was a desire to increase technology education, the actual means lagged.

“Computers were coming to the schools, but they weren't ready. There were computer-using educators that are scattered across the state of the country, but they were really organized together, and they hadn't seen some of the things that we could see coming down the road,” she said. That was, in large part, the impetus for the program, which was then called Foundations for the Future (F3).

She embraced and pioneered the partnership-seeking approach that is now a hallmark of the renamed STEM@GTRI.

Huff started with a small amount of seed funding from GTRI. However, her dogged determination led her to secure $2 million in funding from AT&T to really get the ball rolling. The AT&T funds were leveraged into that all-important funding from the State of Georgia, which continues to the present.

“I think the biggest impact was getting everybody aware, or getting people who needed the resources aware that we have resources, letting them know,” Claudia said.

The principles and practices that she put into place out of necessity became the foundation for what STEM@GTRI is 25 years later.

To honor and thank Claudia Huff, she was presented with STEM @GTRI’s inaugural STEM Champion Award.

Educating Future Technology Leaders

GTRI Director Jim Hudgens said that when he first arrived at GTRI four years ago, STEM @GTRI was one of the first programs he heard about: “I was just blown away by the program,” he said during his opening remarks.

“Educating future technology leaders is one of our core mission areas,” said Hudgens. “A big part of what we do in educating technology leaders is that we take it very seriously. Our people are extremely passionate about this--about their many volunteer hours going out to science fairs, going to high schools across the state, teaching classes in high schools--doing as much as they can.

“It's an amazing community at GTRI that makes this happen.”

That passion and spirit of commitment was noted often during the 25th Anniversary luncheon.

The anniversary event was hosted by Leigh McCook, Director of STEM @GTRI, which she calls “a fun role.” Her passion and commitment to STEM@GTRI was noted by speakers throughout the luncheon program.

“One of the greatest impacts I get to experience is working with our K-12 future STEM workforce. When I see a Georgia Tech/GTRI researcher explain and demonstrate their work to a classroom of elementary, middle, or high school students or experience students of all ages interact with our researchers through questions and discovery — I am thrilled to witness students have that ‘ah ha!’ moment and think ‘This is cool stuff! I want to study to learn to be a (fill in the STEM field here),’ or even ‘Oh, now I know why I’m having to learn this topic in my class — someone really does use this stuff in the real world!’

“When we get to bring diversity to Georgia’s classrooms across the state through our outreach, we open worlds of awareness of possibilities and opportunities for our K-12 students.”

Bringing ‘What If’ to the Real World Through Partnerships

“Real-world” impact, and opening students’ (and teachers’) eyes and minds to possibilities were common themes reiterated by the luncheon speakers.

District 25 State Rep. Todd Jones spoke of several of his “dreams” for the State of Georgia: advancements in daily life, from improved transportation to medical advances—all “dreams” that are dependent on significant advances in technology, which Jones said he believes is incumbent on advancing technology education throughout Georgia, including in rural areas without extensive technology resources or even a large quantity of technology educators. That, he said, is where STEM @GTRI’s outreach is invaluable.

Jones said that his office’s ongoing partnership with GTRI is key to improving the “access and rigor” of STEM education in Georgia.

“I'm going to give all the credit to GTRI. There might have been passion coming out of my office and willingness to find a partner to make this happen, but between Bert (Reeves, Vice President, Institute Relations) and the GTRI team, that is what kind of made this a success.

“We did know that GTRI had the resources to be able to make this work. What they had to deal with for a couple of decades around STEM, around the work, shows a passion and an application. That was what we were looking for.”

McCook noted that Jones’ initiative to improve access and rigor of computer science education across Georgia, as part of the newly funded Rural Computer Science Education Program, shows how committed STEM @GTRI is about fostering and furthering partners. She noted that,  in partnership with Georgia Tech’s Center for Education Integrating Science, Mathematics, and Computing (CEISMC), the project is “in 16 (Georgia school) districts right now” and includes contributions from the Institute for People and Technology (IPaT), the Institute for Robotics and Intelligent Machines (IRIM), and others.

“You can't dream it if you've never been exposed to it,” Jones said enthusiastically “Dreams come from ‘what if,’ but ‘what if’ can't be had unless you know what's possible and maybe what could be next.”

Such a commitment to fostering a sense of making “what if” possible was reiterated by Karen Faircloth, Director of School Improvement & Professional Learning for the Northwest Georgia Regional Education Service Agency (RESA), which encompasses school districts in smaller communities such as Cartersville, Dallas, Rome, and Tallapoosa.

STEM@GTRI High School Internship Program

STEM@GTRI thrives today largely because of the indefatigable efforts of High School Summer Internship Program co-directors Therese Boston, a Senior Research Associate in ICL, and ATAS Principal Research Engineer Erick Maxwell. STEM@GTRI’s High School Internship Program is one of its premier initiatives. In the internship program, Georgia high school students who are at least 16 years old may apply for five-week paid summer internships hosted in GTRI labs. Interns work on projects in GTRI laboratories and the GTRI Warner Robins field office with the goal of providing students with real-world experiences in science and engineering research. GTRI researchers mentor students by working with them on projects to engage them in first-hand STEM experiences.

As an example of the first-hand nature of the internship, Maxwell cited a project done by an intern team in conjunction with the 3rd Infantry Division (3ID) at Fort Stewart, Georgia. The high schoolers developed a means to streamline the arduous task of counting ammunition rounds via the use of “smart” gloves. To further emphasize the tangible benefits of the students’ experience, Maxwell noted that the students are included on the project’s application for a full patent on the gloves.

The High School Internship Program and other programs of STEM@GTRI make use of partnerships with GTRI’s laboratories, Georgia Tech, the U.S. military, and businesses in technology-related industries.

Among the industry representatives in attendance was Patrick Govan, Higher Education Account Manager at Cisco. He explained how his company, a leader in digital communications technologies, works in outreach along with STEM@GTRI. “We are starting to work with the STEM outreach program, bringing some of the students and internships into our office--we just built a new office in the Coda building (at Tech Square). So, we're show showcasing how technology is used in everyday life and in office space to inspire the younger kids. [We show them] a day in the life of what a career would look like in the tech space.

“Leigh (McCook) and I are trying to get the [STEM@GTRI] summer internship program incorporated into office visits and things like that.”

Looking ahead to future goals and activities was very much a part of the 25th-anniversary celebration. Here’s to the next 25 years of STEM@GTRI!

Writer: Christopher Weems 
Photos: Christopher J. Moore
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $940 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

Using a first-of-its-kind automated tool that can assess a website’s password creation policies, researchers also discovered that 12% of websites completely lacked password length requirements.

Assistant Professor Frank Li and Ph.D. student Suood Al Roomi in Georgia Tech’s School of Cybersecurity and Privacy created the automated assessment tool to explore all sites in the Google Chrome User Experience Report (CrUX), a database of one million websites and pages.  

Li and Al Roomi's method of inferring password policies succeeded on over 20,000 sites in the database and showed that many sites:

• Permit very short passwords
• Do not block common passwords
• Use outdated requirements like complex characters

The researchers also discovered that only a few sites fully follow standard guidelines, while most stick to outdated guidelines from 2004. The project was 135 times larger than previous works that relied on manual methods and smaller sample sizes.

More than half of the websites in the study accepted passwords with six characters or less, with 75% failing to require the recommended eight-character minimum. Around 12% of had no length requirements, and 30% did not support spaces or special characters.

Only 28% of the websites studied enforced a password block list, which means thousands of sites are vulnerable to cyber criminals who might try to use common passwords to break into a user’s account, also known as a password spraying attack.

“Both Professor Li and I were excited to take on the challenge,” said Al Roomi. “With his guidance and our continuous work on both algorithm design and the measurement technique, we were able to fully develop an automated measurement of password creation policy and apply it at scale.”

Al Roomi and Li designed an algorithm that automatically determines a website’s password policy. With the help of machine learning, the pair could see the consistency of length requirements and restrictions for numbers, upper- and lower-case letters, special symbols, combinations, and starting letters. They could also see if sites permitted dictionary words or known breached passwords.

“As a security community, we've identified and developed various solutions and best practices for improving internet and web security,” said Li. “It's crucial that we investigate whether those solutions or guidelines are actually adopted in practice to understand whether security is improving in reality.”

The project began during the height of the pandemic when Al Roomi found a gap in the research literature surrounding website password policies. Through his reading, he discovered that a consensus of his peers did not think a large-scale survey of password policies was possible due to the variety of web design.

“It was exciting to see an identified challenge in the literature and to develop and apply a vision we turned into the measurement tool,” said Al Roomi. “This research was my first in my Ph.D. program at Georgia Tech and SCP. It is one of the most challenging yet rewarding endeavors I've worked on.”

The full report will be presented at the ACM Conference on Computer and Communications Security (CCS) in Copenhagen, Denmark, later this month. A Large-Scale Measurement of Website Login Policies was also accepted to the 32nd USENIX Security Symposium earlier this year.

By the end of his demonstration video, Ph.D. student Jason Kim showed how the recently discovered iLeakage side-channel exploit is still a genuine threat to Apple devices, regardless of how updated their software might be.

First discovered by Kim and Daniel Genkin, an associate professor in the School of Cybersecurity and Privacy, the vulnerability affects all recent iPhones, iPads, laptops, and desktops produced by Apple since 2020.

iLeakage allows attackers to see what’s happening on their target’s Safari browser. This vulnerability allows potential access to Instagram login credentials, Gmail inboxes, and YouTube watch histories, as Kim demonstrated last month on a slightly older MacBook Pro.

“A remote attacker can deploy iLeakage by hosting a malicious webpage they control, and a target just needs to visit that webpage,” said Kim. “Because Safari does not properly isolate webpages from different origins, the attacker's webpage is able to coerce Safari to put the target webpage in the same address space. The attacker can use speculative execution to subsequently read arbitrary secrets from the target page.”

How is this possible? Well, as manufacturers developed faster and more efficient CPUs, their devices have become vulnerable to something called speculative execution attacks. This vulnerability is in the design of the chip itself. It has led to major software issues since the Spectre attack was reported in 2018.

There have been many attempts to stop these types of attacks, but Kim and Genkin show through their research that more work still needs to be done.

“iLeakage shows these attacks are still relevant and exploitable, even after nearly six years of Spectre mitigation efforts following its discovery,” said Genkin. “Spectre attacks coerce CPUs into speculatively executing the wrong flow of instructions. We have found that this can be used in several different environments, including Google Chrome and Safari.”

The team made Apple aware of its findings on Sept. 12, 2022. Since then, the tech company has issued mitigation for iLeakage in Safari. However, the researchers note that the update was not initially enabled by default. It was only compatible with macOS Ventura 13.0 and higher as of today.

So far, the team does not have evidence that real-world cyber-attackers have used iLeakage. They‘ve determined that iLeakage is a significantly difficult attack to orchestrate end-to-end, requiring advanced knowledge of browser-based side-channel attacks and Safari's implementation.

The vulnerability is confined to the Safari web browser on macOS because the exploit leverages peculiarities unique to Safari's JavaScript engine. However, iOS users face a different situation due to the sandboxing policies on Apple's App Store. The policies require other browser apps using iOS to use Safari's JavaScript engine, making nearly every browser application listed on the App Store vulnerable to iLeakage.

iLeakage: Browser-based Timerless Speculative Execution Attacks on Apple Devices will be published at the 2023 ACM SIGSAC Conference on Computer and Communications Security later this month.

Along with Kim and Genkin, Stephan van Schaik of the University of Michigan and Yuval Yarom of Ruhr University Bochum co-authored the paper.

Silon US, a Peachtree City manufacturer that designs and produces engineered compounds used to create a wide range of products — from automotive applications to building materials, such as PEX piping and wire and cable, was experiencing problems with their extrusion line during a time of increasing customer demand. Problems with the drive mechanism on that extrusion line, a piece of equipment critical to the company’s ability to produce, threatened to shut them down. With replacement parts several weeks away, was it safe to continue operating? At what throughput rates? How much collateral damage might be incurred if they continued to operate?

That’s when Silon managers turned to GaMEP for help.

After working through ideas with GaMEP’s manufacturing experts, the team installed wireless condition monitoring sensors that provide continuous, real-time insights on their manufacturing assets’ health. With the sensors, Silon was able to find a sweet spot that not only allowed them to continue operating but also kept them from overexerting the equipment, preventing further damage.

The solution to that problem has now become a routine part of Silon’s process, as company technicians continue to use this sensor technology for early detection of any deviations or anomalies in the machinery’s health, allowing the company’s maintenance team to proactively respond by adjusting scheduled maintenance to avoid costly downtime.

GaMEP’s Sean Madhavaraman says, “Silon is more productive than ever and on track for growth. The strong results in this challenge are a great example of the decades-long focus of GaMEP to educate and train managers and employees in best practices, to develop and implement the latest technology, and to work together with businesses to find solutions.”

Daniel Raubenheimer and Matt Gammon, Silon’s general managers, also lauded GaMEP, saying, “GaMEP’s extensive experience within the manufacturing realm has been a great benefit to our company. The wireless condition monitoring sensors allow us to predict future breakdowns and mitigate a potential catastrophe — allowing us to operate in a safe manner, while saving money, time, and effort.”

These areas frequently overlap with abusers often using the internet and mobile technology to extend the reach of their abuse. However, the smaller scale of these online attacks has resulted in less attention from security researchers.

By building on developments recently made in cognitive security, Principal Research Scientist Courtney Crooks and graduate student Sneha Talwalkar are working to bring relief to survivors of domestic abuse.

The impact of domestic abuse, otherwise called intimate partner violence (IPV), on public health is something that Crooks has been studying for several years through research and practice in her role as a licensed psychologist and researcher.

After seeing how new technology opened new methods of abuse online, Crooks realized she could help fill in the gaps in this research space using her experience working with the Georgia Tech Research Institute, the School of Cybersecurity and Privacy (SCP) at Georgia Tech, and the Emory University School of Medicine.

To get what they want, abusers try to change their victim’s state of mind through cognitive manipulation and use different tactics to do so. Crooks decided to explore ways to help IPV survivors counteract these enhanced technology-enabled cognitive security risks as they progressed through their recovery.

The software Crooks and Talwalkar are working to develop would alert survivors to these potential or observed abuses by leveraging well-known, developmentally appropriate, psychologically based learning strategies. The tool will focus solely on the unique risks faced by IPV survivors. Applying human-centered design principles and ethical standards to the AI design will be a top priority for the team.

The team is working to develop AI-assisted interventions that are psychologically informed and made specifically to focus on the unique risks faced by survivors. These interventions will be designed to take place alongside traditional methods of support, such as mental health and community resources.

“It’s important to understand that abusive relationships are complicated. While some people can escape them, many can’t,” said Crooks. “Or they may physically escape, but resources like their phones, online accounts, or finances may still be vulnerable to their abusers. Survivors may also need to continue to communicate with their abuser, like in instances in which they share children.”

Regardless of circumstances, it is often difficult for survivors to stop communicating with their abusers once they escape the relationship. This inability to disconnect is because of the psychological connections reinforced while they were with their former partner.

The AI technologies Crooks and Talwalkar propose will not act like a ChatGPT chatbot. Instead, it will act like a coach, learning from abusive behavior tactics and potential survivor responses.

The tool will then make suggestions based on each user’s specific recovery progress and goals while factoring in potential risks. To improve its coaching performance and general knowledge base, the AI will continue to learn from the outcome of each incident survivors face.

“The model provides the necessary intervention to assist in the recovery of an IPV survivor,” said Talwalkar. “We want to use artificial intelligence for good, and this project is a step in that direction.”

The classes in the SCP master’s program played a pivotal role in shaping Talwalkar’s research in this area. While exploring internet censorship and language models, she recognized the emerging challenges posed by AI in security. After an insightful conversation with SCP Professor Peter Swire, Talwalkar gained the confidence to shift her focus towards investigating malicious intent in immersive environments. With Crooks’ guidance, she began exploring the socio-technical environment of IPV.

Designing User-Centered Artificial Intelligence to Assist in Recovery from Domestic Abuse was accepted as an extended abstract and presented to the 2023 World Congress Computer Science, Computer Engineering, and Applied Computing event this summer. Proceedings of the IEEE is publishing the work in an upcoming issue.  

In May, Crooks, Talwalkar, and others from their research team presented their findings at the Health Sciences Research Day hosted on the Emory University campus by the Emory School of Medicine. Crooks presented her study of the lived experience of coercive control in domestic abuse, from which this current research is derived, at the February 2023 National Meeting of the American Psychoanalytic Association. 

October is National Domestic Violence Awareness Month and National Cybersecurity Awareness Month. For more information about domestic abuse and resources to help, please visit the Centers for Disease Control and Prevention website.

This year’s theme will center on AI and responsible stewardship. It will feature keynote speakers, breakout sessions, panel discussions, networking opportunities, and an expo showcasing research and real-life applications of AI.​ The goal is to create an experience for all levels of professionals and entrepreneurs where they can come to Georgia Tech and interact with speakers, researchers, and industry leaders.

​Topics will include AI in education, healthcare, arts and culture, and industry, as well as the role of regulatory and governmental oversight. At the Street Innovation Showcase, located in Tech Square and the Coda courtyard, students and faculty from universities and colleges across metro Atlanta, along with corporations, will present their research and applications of AI. The showcase is free and open to the public.

Avant South is part of InnovATL, a citywide platform that amplifies the voices of metro Atlanta entrepreneurs, innovators, and creators. This year’s edition is scheduled to take place during the first week of a monthlong celebration of Atlanta’s innovation ecosystem — between Venture Atlanta, one of the nation’s largest venture capital conferences, and the AC3 Festival, which honors hip-hop culture and the city’s impact in music, technology, and entertainment.  

The event kicks off the evening of Thursday, Sept. 28, with Southern Hospitality, a dinner on the Coda rooftop. Charles Isbell, ICS 1990, former dean of Georgia Tech’s College of Computing and currently the provost and vice chancellor for Academic Affairs at the University of Wisconsin-Madison, will be the featured speaker. 

On Friday, Sept. 29, César A. Hidalgo, director of the Center for Collective Learning at the Artificial and Natural Intelligence Institute at the University of Toulouse, will be the morning keynote speaker. The lunch keynote is Kathy Baxter, PSY 1996, M.S. PSY 1998, and principal architect of responsible AI and tech at Salesforce. There will be a fireside chat with Tom Gruber, co-founder and head of design for the company that created Siri, to wrap up Avant South. Rose Scott from 90.1 WABE will be a host and moderator throughout day two, which will take place at the Fox Theatre. 

“Avant South will showcase Georgia Tech as a trusted public voice that informs decision-makers in business, academia, and public policy on issues of consequence,” said President Ángel Cabrera. “We want to cement our role as the go-to institution for finding solutions for global challenges, and I can’t think of a more timely or urgent topic than managing the impact of AI on society and making ethical and responsible decisions about how to use and create this technology.”

For more information, including a list of top presenters, visit avantsouth.com.

Findings produced by Georgia Tech's Cyber Forensics Innovation (CyFI) Lab reveal this new type of menace, labeled as web-app-engaged (WAE) malware by the lab, has seen an increase of 226% since 2020. Fortunately, the team created a tool that enables cybersecurity incident responders to purge nearly 80% of discovered WAE malware by collaborating with service providers. 

“Web applications have become an integral part of our online lives, offering various services such as content delivery, data storage, and social networking,” said Mingxuan Yao, Georgia Tech Ph.D. student. “Unfortunately, these utilities have made web applications an attractive playground for malware creators. WAE malware is designed to exploit these applications, posing several risks to users.” 

WAE malware operates deceptively, though not in the ways one might expect. Rather than compromising the security of the web applications, this type of malware abuses the applications by making its malicious traffic appear benign. By doing so, it effectively hides in plain sight, enabling it to carry out its activities without being detected. 

Addressing these threats requires a coordinated effort between incident responders and web app providers. Still, such collaboration has been lacking until now. The research produced by CyFI Lab seeks to enable such cooperation and provide insights into the prevalence and the characteristics of WAE malware. 

Yao and his co-authors created Marsea to comprehensively examine WAE malware automatically. The tool identifies and separates abuse based on a web app’s identity and assets. 

When used on a group of 10,000 malware samples, Marsea found nearly a thousand instances of malware throughout 29 different web applications. Alarmingly, Marsea also revealed that attackers are transitioning their malicious command-and-control servers to these web apps to evade detection. The research team has used Marsea to collaborate with web app providers to take down 79.8% of the malicious web app content. 

In August, the team presented Hiding in Plain Sight: An Empirical Study of Web Application Abuse in Malware at the 32nd USENIX Security Symposium. Jonathan Fuller of the United States Military Academy, Georgia Tech Ph.D. students Ranjita Pai Kasturi, Saumya Agarwal, Amit Kumar Sikder, and Assistant Professor Brendan Saltaformaggio co-authored the paper. 

Online ads like this are all too familiar and often the opening salvo in personal cyberattacks that can lead to unwanted software or other malicious downloads.

Georgia Tech researchers are countering deceptive online ads with a pioneering solution designed to challenge the rising threat of online social engineering attacks by cutting them off at the source.

Trident, created by Ph.D. student Zheng Yang and his team of researchers, is an add-on compatible with Google Chrome that has proven to block these ads with nearly 100% efficiency.

Advertisements are fertile ground for scams and fraudulent schemes. While such networks may offer better pay to websites than industry giants like Google and Facebook, their advertisements often employ tactics that lure unsuspecting users into compromising situations.

“The goal is to identify suspicious ads that often take users to malicious websites or trigger unwanted software downloads,” said Yang. "Trident operates within Chrome’s developer tools and uses a sophisticated AI to assess potential threats."

The team compiled a vast dataset from over 100,000 websites to build Trident, including ten low-tier ad networks. This comprehensive data collection helped identify 1,479 instances of attacks encompassing a range of six common types of web-based social engineering attacks. These include:

• Tech-support scams
• Unwanted software downloads
• Scareware
• Dating scams
• Notification spam
• Prize scams

The remarkable outcome of their efforts is the sustained performance of Trident. Over the course of a year, the tool consistently achieved a nearly perfect detection rate of malicious ads, ensuring users' safety by minimizing the risk of interacting with harmful content.

Impressively, this achievement came with a meager 2.57% false positive rate, demonstrating the accuracy and effectiveness of Trident's machine-learning capabilities.

TRIDENT: Towards Detecting and Mitigating Web-based Social Engineering Attacks was presented at the 32nd USENIX Security Symposium in August. Contributors to this project include Georgia Tech Ph.D. students Joey Allen and Matthew Landen, Adjunct Assistant Professor Roberto Perdisci, and Professor Wenke Lee.

That’s true at home and at work. And it’s true for the critical systems the U.S. Department of Defense relies on to protect the nation.

Think about all the highly sophisticated systems that power drones or fighter jets or even secure authentication programs. Many of those systems are custom software developed at great expense. Which means updating them isn’t as easy as downloading the latest software patch and clicking “Install.”

It often requires a time-consuming rewrite or reverse engineering process that costs even more time and money. But not if a team of Georgia Tech engineers and cybersecurity researchers are successful. They’re among the teams working to speed up the process with a $10 million Defense Advanced Research Projects Agency (DARPA)-funded effort to unpack these legacy systems, incorporate updates, and redeploy them in weeks or months rather than years.

“The U.S. government has this tremendous problem where they put tons of research and development into cutting edge software, and then two years down the line, it needs to be updated or applied to a new platform or it needs patches. We can’t just go back to the drawing board and rewrite all of our software every few years,” said Brendan Saltaformaggio, an associate professor in the School of Cybersecurity and Privacy (SCP) and the School of Electrical and Computer Engineering (ECE).

Read about the team's work on the College of Engineering website.

Nearly 250 attendees from more than 200 government, academic, and industry organizations convened at the Phoenix Challenge June 20-23 to discuss how misinformation, disinformation, and the propagation of bad information may affect the world – and how organizations across those three sectors can work together to address growing concerns about the effects of what’s happening in this arena. Although AI was among the top concerns, there were many other issues on the agenda.

The conference was organized for the Office of the Undersecretary of Defense for Policy (OUSDP) by GTRI, the University of Maryland Applied Research Laboratory for Intelligence and Security (ARLIS), and the Information Professionals Association.

The June Phoenix Challenge conference was part of a series of events designed to promote collaboration on efforts ranging from research and acquisition to operational planning and execution, with goals of reducing enterprise ambiguity in the Department of Defense, promoting awareness, and exchanging information. Recommendations coming out of the meeting’s working groups are being briefed to appropriate offices in the Department of Defense and other agencies.

“The idea for the Phoenix Challenge is to create a watering hole where everyone can participate with equal standing,” said Austin Branch, professor of the practice at ARLIS, which is funded by the OUSDP to convene the Phoenix Challenge events. “By bringing these communities together, government can enjoy additional critical thinking and testing of ideas, offering new concepts, technologies, and methodological approaches in an environment that’s collaborative and includes everyone.”

OIE – a discipline that in years past was known as information warfare – can include such topics as electronic warfare, cyber operations, military deception, and psychological operations in a broad cognitive security space. “The Phoenix Challenge is a recognized platform for collaboration and sharing, in both technical and non-technical areas, and in the hard sciences and soft sciences,” Branch said. “Participants have to be prepared to work because we’re working on solutions, and there is a sense of mutual accountability.”

Beyond the recommendations to the government, participants from industry and academic communities benefit from obtaining a better understanding of the government’s needs, plans, and concerns.

“Here, we can have everybody concentrated and focused, with a great value proposition in being able to reduce ambiguity about what the requirements are and for the government to articulate what the needs are, then allow this broader enterprise to work on those things,” Branch added.

At the Atlanta meeting, there were three panel discussions, including one on generative AI, which has both positive and negative implications for the world’s information environment.

“This technology is going to have an enormous impact on us going forward,” said Theresa Kessler, a GTRI research scientist who was among the Atlanta event’s organizers. “AI and machine learning tools can make the OIE challenges worse, or be used to make them better. There’s also a cybersecurity component and the human element of how people can be so accepting of bad information.”

The goals of the Phoenix Challenge include much more than identifying the issues. Attendees participated in six working groups organized to highlight potential solutions and make recommendations to be considered by the government. And those making the recommendations are expected to play a role in carrying them out.

“Ultimately, the goal is to affect the national defense strategy, with these output products, recommendations that the working groups built,” Kessler explained. “We had a huge representation of industry partners, along with academic participants, including multiple universities, University Affiliated Research Centers (UARCs), and Federally-Funded Research and Development Centers (FFRDCs). Each of our working groups had a representation from industry, government, and academia.”

That broad representation helped provide a perspective not limited to a single constituency, she said. “The working groups were designed and facilitated in a way that everybody’s opinion was pulled in and valued. Involving all these different groups provides a more holistic presentation of the problem and the solution set.”

In addition to a classified working group, the breakout sessions focused on:

• Inputs to the R&D Roadmap for OIE Technologies.
• Detection and Beyond: Implementing Effective Technological Solutions to Emerging OIE Threats.
• Applied Research: Assessments.
• Strategy for Operations in the Information Environment (SOIE) Implementation Plan Framework.
• Resilience to Adversary Disinformation.

Among the conference speakers were:

• Todd Breasseale (Deputy Assistant to the Secretary for Public Affairs, Office of Information Operations Policy).
• LtGen (R) Dennis Crall, USMC.
• Heidi Shyu, Under Secretary of Defense for Research and Engineering (OUSD(R&E)), who addressed the conference virtually.
• Neill Tipton, Director for Defense Intelligence, Collection and Special Programs.

The June Phoenix Challenge event was the first hosted by GTRI, but the event has a long history, beginning decades ago and including recent meetings in London and Charleston, South Carolina. In 2022, GTRI hosted an Information Warfare Summit on its Atlanta campus, but elected to join forces with the Phoenix Challenge in 2023. The next event is likely to be held in the Washington, D.C., area during 2024.

Writer: John Toon (john.toon@gtri.gatech.edu)
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $800 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

Ayush Pai, an avid NBA viewer, set out to create a solution as he watched fouls go uncalled and outcomes be decided by officials rather than the players.

With a self-taught knowledge of computer vision AI, he created the initial model for the official­­ –– dubbed version 1.0 –– in the summer of 2022 after graduating high school. With soccer's video assistant referee (VAR) system and the Hawk-Eye review system in tennis as a guide, Pai felt that basketball was missing an AI component to supplement referees on the court.

"The officials are running up and down the court constantly. Basketball in general, especially the NBA, is one of the most fast-paced sports, and the entire court is difficult for officials to see at all times. Why not implement a system that can enhance their ability to see the floor using multiple cameras with real-time processing and detection of the basketball and the players?" he said.

Version 1.0 of Pai's system was trained to detect the color of the basketball and used a pedometer to track steps, but lighting issues and the use of a cumbersome step counter proved limiting. In June, the Denver Nuggets capped a 4-1 series victory over the Miami Heat to earn their first NBA championship, but officiating mistakes in the Heat's lone win propelled Pai to get back to work on his concept. By the end of the month, version 2.0 of the AI official was complete.

To improve accuracy, he used over 3,000 photos of basketballs to better train the model to track the ball's movement, and he also used pose estimation for tracking the position of a player's ankles and wrists to detect double dribbling and traveling violations. He is now working on version 3.0, which will include the capacity for multiple players, cameras, and the ability to detect shooting fouls and reach-ins. Pai is also developing a mobile app that he hopes will bring accurate officiating to youth leagues and pickup games worldwide.

The AI craze shows no signs of slowing down, and Pai sees his creation as a way to introduce the concept to new audiences. "Everyone should see the potential of AI, and it's important that everyone has a basic understanding of its implications and how it can be used in different fields. You can use AI for ChatGPT, in Tesla's autonomous systems, and even in sports," he said.

An ongoing debate amid the rise of AI is its potential to displace people. Pai sees the benefit of his system as one of coexistence rather than a replacement for the human element of officiating.

"You don't want to make AI feel like a threat to the people in any industry. It's a much better approach to consider this a partnership between AI and officials. In basketball, there are so many physical aspects of the job that AI can't do," he said.

The Florida native will trade his University of Florida Gator blue and orange for Yellow Jacket white and gold this fall as a transfer student, and the incoming second-year student is eager to get started.  

"One of the main reasons I transferred is because of the opportunities the Institute offers within AI and computer science,” he said. “The chance to participate in the College of Computing's Threads program will help me with my projects, and I’m looking forward to exploring the entrepreneurial side of Tech with great programs that foster innovation." 

Pai’s interest in the capabilities of AI isn’t limited to basketball. He has also created systems that attempt to discourage texting and driving and keep students on task while studying.

Georgia Tech continues to cement itself as a leader in the emerging space, highlighted by the recent launch of the AI Hub on campus, which will drive education and research while developing real-world, responsible applications for AI.

Sanders found the perfect fit in GTRI’s Professional Education program (GTRI-PE). GTRI-PE is connected to the Georgia Tech Professional Education program (GTPE) and offers short courses and certificate programs taught by GTRI researchers in the areas of defense technology, cybersecurity, and occupational safety and health.

Sanders last year earned a cybersecurity certificate through the program, which teaches participants how to best mitigate risk, defend their organization from external and internal threats, and more. Working in a cyber-focused role at GTRI, Sanders said the program equipped her with the strategic and technical knowledge to help protect GTRI against emerging threats.

“The program was exactly what I was looking for,” said Sanders. “It fit into my schedule and helped me obtain more career-specific credentials.”

GTRI-PE offers over 100 distinct courses taught by more than 160 instructors. During FY22, the program delivered a total of 184 courses, predominantly catering to organizations such as the U.S. Department of Defense (DoD) and various government sponsors. GTRI researchers with a suitable background and proficiency may serve as instructors. Instructors receive supplemental compensation as an acknowledgment of their contributions to the program.

GTRI-PE Director Renita Folds said GTRI researchers provide a practical perspective to the classroom that extends beyond theories and concepts.  

“GTRI researchers bring immense value to our short courses, primarily through their extensive experience in their respective fields,” Folds said. “They are actively engaged in applied research and working on cutting-edge solutions for complex problems on a daily basis. This direct involvement in the field allows them to bring real-world insights and up-to-date knowledge to the classroom, enhancing the learning experience for our course participants.”

GTRI places high importance on providing courses that cater to the current demand in various fields. While radio frequency (RF) electromagnetic warfare (EW) and cybersecurity remain highly sought-after disciplines, the Georgia Institute of Technology (Georgia Tech) also recognizes the significance of emerging technologies. Hence, GTRI is prioritizing the development of courses focused on cutting-edge subjects like artificial intelligence (AI), machine learning (ML), and data science. A new communications certificate program is currently under development and is set to launch in FY24, said Folds.

“We recognize how critical it is for our government and industry partners to stay ahead of these pressing issues,” she said.   

GTRI-PE offers a mix of in-person, hybrid and virtual classes, which consist of lectures, discussion sessions, and hands-on projects.  

While instructors are considered to be experts on the topics that they teach about, GTRI Principal Research Engineer Carlos Dávila, who teaches courses on radar systems and electronic warfare (EW), said he is often just as much a student as a teacher.

Dávila has been an instructor for the past 20 years, and developed two short courses for the program – Modeling and Simulation of Radar Systems and Basic Electronic Warfare Modeling, which are centered on two widely-used programming languages, MATLAB and Simulink.  

“The intent with these courses is to build upon theoretical concepts by having students develop models that reinforce and illustrate those fundamentals,” Dávila said.

Dávila said his favorite part of being an instructor is gaining fresh perspectives from students, who help him stay current on the ever-changing dynamics of his field.

“I see teaching and performing research as very complementary,” he said. “My students keep me hungry to improve both the breadth and depth of my knowledge base.”

Another instructor, GTRI Principal Research Scientist Matt Guinn, has also been with the program for 20 years and developed the cybersecurity course Introduction to Penetration Testing. Guinn’s class is lab-based and provides students with an understanding of the fundamental threat vectors and exploitation techniques adversaries use to breach systems and networks.

Guinn also co-teaches a course related to his own class called Defensive Cyber Operations. This course is also lab-based and introduces students to modern defensive skills required to counteract cyber threats.

“I often teach these courses back-to-back, which is fun because students get to spend the first class thinking about threats from an adversary’s perspective, and then flip things around and learn about how to best defend against those threats in the second class.”  

Guinn most enjoys demystifying cyber threats and providing his students with practical tools to be prepared to defend their organizations against them.  

“One of the main things that I try to accomplish with my class is to teach professionals who may have a limited amount of technical experience with handling cyber breaches the fundamentals of how to best address them,” he said.

But GTRI-PE is not limited to novices.  

Industry veterans who participate in the program say they can’t believe how much there is left to learn.

Jaime Downing, an information security manager at the Naval Air Systems Command (NAVAIR), Naval Air Warfare Center Aircraft Division (NAWCAD), which provides integrated air warfare capabilities to the U.S. Navy, has close to 25 years of cybersecurity experience, an MS in Information Systems Management Cybersecurity and multiple cyber certifications. Downing has audited cyber courses offered by similar programs across the country.

Downing, who earned GTRI’s cybersecurity certificate in 2021, said the practicality of the classes and the ability to collaborate with other DoD professionals helped her view cyber concepts in a new light.  

“GTRI provided perspectives to help with delivering objectives and benchmarks associated with vulnerabilities, threats and risk reduction,” Downing said. “Even at the expert level, there is something new to learn every day. The GTRI team provided professionalism and friendliness, displayed significant details, and was well versed on the topics taught.”

Downing added that the program reinforced the importance of maintaining strong cyber networks from a national security standpoint.  

“From a DoD perspective, the U.S. has to be trained one step further than its adversaries,” she said. “We need to make sure that we are as cyber-savvy as possible and that all of our networks are secured. Our nation’s future depends on it.” 

If you are interested in learning more about GTRI-PE, you may contact Renita Folds at renita.folds@gtri.gatech.edu.

Writer: Anna Akins 
Photos: Sean McNeil 
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $800 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

The ability to uniformly scale the width (number of neurons) and depth (number of neural layers) of a DNN model means that remote clients can equitably participate in distributed, real-time training regardless of their computing resources. Resulting benefits include improved accuracy, increased efficiency, and reduced computational costs.

Developed by Georgia Tech researchers, the ScaleFL framework advances federated learning, which is an ML approach inspired by the personal data scandals of the past decade.

Federated learning (FL), a term coined by Google in 2016, enables a DNN model to be trained across decentralized devices or servers. Because data aren’t centralized with this approach, threats to data privacy and security are minimized.

The FL process begins with sending the initial parameters of a global DNN model to smartphones, IoT devices, edge servers, or other participating devices. These edge clients train their local version of the model using their unique data. All local results are aggregated and used to update the global model.

The process is repeated until the new model is fully trained and meets its design specifications.

Federated learning works best when remote clients involved in training a new DNN model have comparable computational power and bandwidth. But training can bog down if some participating remote-client devices have limited or fluctuating computing resources.

“In most real-life applications computational resources tend to differ significantly across clients. This heterogeneity prevents clients with insufficient resources from participating in certain FL tasks that require large models,” said School of Computer Science (CS) Ph.D. student Fatih Ilhan.

“Federated learning should promote equitable AI practice by supporting a resource-adaptive learning framework that can scale to heterogeneous clients with limited capacity,” said Ilhan, who is advised by Professor Ling Liu.

Ilhan is the lead author of ScaleFL: Resource-Adaptive Federated Learning with Heterogeneous Clients, which he is presenting at the 2023 Conference on Computer Vision and Pattern Recognition. CVPR 23 is set for June 18-22 in Vancouver, Canada.

Creating a framework that can adaptively scale the global DNN model based on a remote client’s computing resources is no easy feat. Ilhan says the balance between a model’s basic and complex feature extraction capabilities can be easily thrown out of whack when manipulating the number of neurons or the number of neuron layers of a DNN model.

“Since a deeper model is more capable of extracting higher order, complex features while a wider model has access to a finer resolution of lower-order, basic features, performing model size reduction across one dimension causes unbalance in terms of the learning capabilities of the resulting model,” said Ilhan.

The team overcomes these challenges in part by incorporating early exit classifiers into ScaleFL.

These ML-based tools are designed to optimize accuracy and efficiency by introducing intermediate decision points in the classification process. This capability enables a model to complete an inference task as soon as it is confident in its prediction, without having to process the whole model.

“ScaleFL injects these classifiers to the global model at certain layers based on the model architecture and computational constraints at each complexity level. This enables forming low-cost local models by keeping the layers up to the corresponding exit,” said Ilhan.

“Two-dimensional scaling with splitting the model along depth and width dimensions yields uniformly scaled, efficient local models for resource-constrained clients. As a result, not only does the global model achieves better performance compared to baseline FL approaches and existing algorithms, but local models at different complexity levels also perform significantly better for clients that are resource-constrained at inference time.”

The exit classifiers that help balance a model’s basic and complex features also play into the second part of ScaleFL’s secret sauce, self-distillation.

Self-distillation is a form of knowledge distillation, which has been used to transfer knowledge from a ‘teacher’ model to a smaller ‘student’ model. ScaleFL applies this process within the same network by comparing early predictions made by the exit classifiers (students) and the final predictions of the last exit (teacher) of local models during optimization. This technique prevents isolation and improves the knowledge transfer among subnetworks of different levels in ScaleFL.

Ilhan and his collaborators extensively tested ScaleFL on three image classification datasets and two natural language processing datasets.

“Our experiments show that ScaleFL outperforms existing representative heterogeneous federated learning approaches. In local model evaluations, we were able to reduce latency by two times, and the model size by four times, all while keeping the performance loss below 2%,” said Ilhan.

One paper found that adding visual attributes to text in multimodal models could boost performance and usefulness for humans.

Another study determined that few-shot learning (FSL) models lack robustness against adversarial treatments and need improvements.  

Georgia Tech Assistant Professor Srijan Kumar and Ph.D. student Gaurav Verma lead the research being presented at the upcoming 61st Annual meeting of the Association for Computational Linguistics (ACL 2023).

Co-authors from Georgia Tech joining Kumar and Verma include Shivaen Ramshetty and Venkata Prabhakara Sarath Nookala, as well as Subhabrata Mukherjee, a principal researcher at Microsoft Research.

ACL 2023 brings together experts from around the world to discuss topics in natural language processing (NLP) and AI research. Kumar’s group offers to those discussions their work that focuses on robustness in AI models. 

“Security of AI models is paramount. Development of reliable and responsible AI models are important discussion topics at the national and international levels,” Kumar said. “As Large Language Models become part of the backbone of many products and tools with which users will interact, it is important to understand when, how, and why these AI models will fail.”

[MICROSITE: Georgia Tech at ACL 2023]

Robustness refers to the degree to which an AI model’s performance changes when using new data versus training data. To ensure that a model performs reliably, it is critical to understand its robustness. 

Trust is of essential value within robustness, both for researchers that work in AI and consumers that use it.

People lose trust in AI technology when models perform unpredictably. This issue is relevant in the ongoing societal discussion about AI security. Investigating robustness can prevent, or at least highlight, performance issues arising from unmodeled behavior and malicious attacks.

Deep Learning for Every Kind of Media

One aspect of AI robustness Kumar’s group will present at ACL 2023 delves into multimodal deep learning. Using this method, AI models receive and apply data through modes ranging from text, images, video, and audio.

The group’s paper presents a way to evaluate multimodal learning robustness called Cross-Modal Attribute Insertions (XMAI). 

XMAI found that multimodal models perform poorly in text-to-image retrieval tasks. For example, adding more descriptive wording in search text for an image, like from “girl on a chair” to “little girl on a wooden chair,” caused the correct image to be retrieved at a lower rank.

Kumar’s group determined this when XMAI outperformed five other benchmarks in two different task retrieval tests.

“By conducting experiments in a sandbox setting to identify the plausible realistic inputs that make multimodal models fail, we can estimate various dimensions of a model’s robustness,” said Kumar. “Once these shortcomings are identified, these models can be updated and made more robust.” 

Labels Matter When It Comes to Adversarial Robustness

Prompt-based few-shot learning (FSL) is another class of AI models that, like multimodal learning, uses text as input.

While FSL is a useful framework for AI to improve task performance when labeled data is limited, Kumar’s group points out in their ACL findings paper that there is limited understanding of the methods’ adversarial robustness.  

“Our findings shine a light on a significant vulnerability in FSL models – a marked lack of adversarial robustness,” Verma explained. “This indicates a non-trivial balancing act between accuracy and adversarial robustness of prompt-based few-shot learning for NLP.”

Kumar’s team ran tests on six GLUE benchmark tasks, comparing FSL models with fully fine-tuned models. Here, they found a notable, greater drop in task performance of FSL models treated with adversarial perturbations than that of fully fine-tuned models. 

In the same study, Kumar’s group found and proposed a few ways to improve FSL robustness.

These include using unlabeled data for prompt-based FSLs and expanding to an ensemble of models trained with different prompts. The group also demonstrated that increasing the number of few-shot examples and model size led to increased adversarial robustness of FSL methods.

“Improved adversarial robustness of few-shot learning models is essential for their broader application and adoption,” Verma said. “By securing a balance between robustness and accuracy, all from a handful of labeled instances, we can potentially implement these models in safety-critical domains.”

The dataset compiles meeting minutes, speeches, and press conference transcripts from the Federal Open Market Committee (FOMC). It is the largest tokenized and annotated dataset of its kind.

An investment strategy developed using the dataset predicted investment returns yielding 163.4% higher than the buy and hold method on the QQQ index fund from 2011 to 2022.

The dataset and strategy are part of new research findings from Georgia Tech. The findings document the influence the FOMC has on markets and the economy through its public communications. The research is being presented this month at the 61st Annual Meeting of the Association for Computational Linguistics (ACL 2023).

“By understanding the impact of FOMC communications on market movements, investors can make more informed decisions, and potentially protect their portfolios from sudden downturns or capitalize on growth opportunities,” said Ph.D. student and lead researcher Agam Shah.

“Additionally, it can help economists at the Federal Reserve Banks more efficiently understand the impact of their communication.”

[MICROSITE: Georgia Tech at ACL 2023] 

The dataset contains 214 meeting minutes, 1,026 speeches, and transcripts from 63 press conferences. The meeting minutes and speeches span from January 1996 to October 2022. The press conference archive dates from April 2011 to October 2022.

To explore this heap of FOMC pronouncements, Shah and his team crafted a novel machine-learning classification task. The new task categorized statements in the dataset as hawkish, dovish, or neutral, rather than just positive, negative, or neutral.

The classification task allows computer models to understand FOMC policy stances through the language used in their correspondence. This in turn guides models to predict how markets react to communications, giving investors valuable information to form their own strategies. 

“One of the reasons our research achieved these remarkable results is because it harnesses the power of natural language processing (NLP) to systematically analyze a vast amount of data which is impractical for humans to process effectively,” Shah said. “This provides a much more nuanced understanding of the market’s response to FOMC communications.”

Shah is a Ph.D. student in the School of Computational Science and Engineering (CSE). He is advised by Sudheer Chava, a professor in the Scheller College of Business. Suvan Paturi, a Georgia Tech alumnus and software engineer at Nasdaq eVestment, co-authored the paper with Shah and Chava.

The group will present their paper at a time when the FOMC and the Federal Reserve are in news headlines now more than ever. To curb inflation, the Fed has increased interest rates ten consecutive times from March 2022 to June 2023.

One example that inspired the group occurred during this period on Aug. 26, 2022. Here, FOMC Chair Jerome Powell gave an eight-minute speech that resulted in an almost $3 trillion decline in U.S. equity market value.

This study not only affirms that markets are reactive to words spoken through public communications but now those effects can be measured and predicted. It also provides new tools to help investors make better, more informed decisions. 

“The application of computational methods to finance and economics revolutionizes the way analysts interpret data. It enables us to handle enormous datasets and extract valuable insights that would otherwise remain hidden,” Shah said. 

“This empowers decision-makers to craft strategies that are based on a deeper understanding of market dynamics, leading to potentially higher returns and more efficient financial systems.”

Georgia Tech space policy expert Mariel Borowitz thinks she has a way to help clear up some of that confusion. Under a new $1.3 million grant from the U.S. Department of Defense, Borowitz plans to help lead a major series of public space wargaming exercises. They’re meant to tease out how current U.S. deterrence strategies might fall short when it comes to stopping a conflict in space and what can be done to improve them.

“When it comes to conflict in space, the stakes are enormously high and the challenges are extremely complex,” said Borowitz, an associate professor in the Sam Nunn School of International Affairs, a unit of the Ivan Allen College of Liberal Arts. “This project will better equip us to understand whether existing deterrence models can help hold the line in space or whether another model is necessary to prevent a potentially devastating outbreak in orbit.”

Jon Lindsay, an associate professor in the Nunn School with a joint appointment in the School of Cybersecurity and Privacy, will work with Borowitz on the project, as will U.S. Space Force Lt. Col. Brian Stewart — a Nunn School Ph.D. graduate who now teaches at the U.S. Air Force Academy. Jacquelyn Schneider — a Hoover Fellow at The Hoover Center at Stanford University — rounds out the team.

A central theme of the project will be trying to understand how the concept of integrated deterrence applies to conflict in space. Integrated deterrence essentially boils down to a country using everything at its disposal to prevent conflict from escalating too far, from applying diplomatic and economic pressure to bringing the military into the mix.

Using such means to deter conflict in a global hotspot on the ground is tricky enough. Look no further than Ukraine for contemporary evidence of that.

But when that hotspot is space, conflict doesn’t just threaten stability in one part of the planet. It could quickly become a serious threat to civilian communications, commerce, and military operations across the globe. Despite the high stakes, trying to understand how to tamp down such conflict is something government officials and scholars are only beginning to tackle.

Much of the work in this space focuses on improving military technology to sense what adversaries are doing and improving the ability of militaries to destroy incoming attacks quickly. But this project highlights how no complex problem can be solved without considering both technological and human factors — a core competency of the Nunn School and the Ivan Allen College of Liberal Arts.

“We understand entanglement from a technological standpoint, but we need to better understand how these entanglements affect perceptions and decisions, which ultimately shape deterrence,” Borowitz said. “And we need to have more clarity on how decisions to separate military and civilian systems or choices to integrate different sectors within the space domain more closely might affect deterrence, before billions of dollars are spent on these efforts.”

Borowitz and her colleagues have already staged versions of space conflict scenarios in the classroom at Georgia Tech. They are now broadening the scope and preparing for the first exercises, which could come as soon as September.

The team plans to hold wargaming sessions across the globe over the next few years, including at Georgia Tech and the Air Force Academy and in Washington, Brussels, Taiwan, and Tokyo. The sessions will include national security figures, scholars, students, and international partners.

The project is expected to generate a significant dataset of use to scholars, as well as a book, game design materials, and other assets to help other researchers continue the work, Borowitz said

In a digital age, an endless amount of data is available at the tap of a screen. That accessibility allows Doug Sims to keep all 482 of Infrastructure and Sustainability's (I&S) maps up to date for the Georgia Tech community.

Sims arrived at Tech in 2008 as a utility analyst. With a background in civil engineering and the keen eye of a land surveyor, he began identifying ways to streamline operations using a geographic information system (GIS). At the time, Sims explained that GIS was seldom used outside of tax assessors’ offices, but he recognized its ability to connect lines on a page to valuable data.

"If you're looking at a line for a pipe, you can click on the pipe and see that it's a 10-inch pipe, and you can see what it carries, what it's used for, and where it goes," Sims said.

Over the past 25 years, Sims noted how GIS has exponentially evolved to map out entire countries. Georgia Tech is often described as a city within a city, and with the help of GIS, that presents an opportunity for the department to get a better lay of the land.

"We have our own electrical distribution system. We have our district energy for the chilled water and the steam managed by I&S. When you look at how GIS is normally used, it's normally looking at larger scales. So, once you start shrinking it to a relatively small area like the Tech campus, it changes how you look at things as they become much more detailed,” he said.

Now the senior systems support engineer for Infrastructure and Sustainability, Sims' foresight to increase GIS application was a catalyst for the data sets that exist today. In his role, Sims continues to search for ways to capture additional data points and recognizes that the communal nature of a campus setting can provide opportunities for instantaneous collaboration.

"It's a big paradigm shift. It's putting power in the hands of the people where they can make the changes immediately. Whether I'm sitting on campus or 50 miles away, I can make a change. They can hit refresh and see that change immediately. It's changing how we think about and use that data every day," he said.

A recent project, working alongside the recycling and zero waste department, invited students to identify the location of all trash cans and recycling bins on campus. Sims' creation allowed participants to use their phones to pinpoint locations that were added to the map, which has since been used for additional analysis of the Institute's sustainability efforts.

The ability to connect with data instantaneously has helped Sims and his staff, which recently added a new team member –– a Tech alumna who worked with the department as an undergraduate –– to expedite work orders across campus. The first widespread application of GIS mapping empowers members of the Tech community to report landscaping requests to I&S by marking the exact location of the issue, sending photos and any other relevant information directly to an organized dashboard for review. Sims hopes to expand this service to additional services in the coming years.

Keeping campus informed and operating at its most efficient is a point of personal pride for Sims, and, wielding what he calls the “GIS hammer,” he hopes to be a springboard to assist other departments using GIS.

"If people want to map something or see what we have, get in touch with me, and let's see if there's a solution we can provide to you," he said. "I'm here to come up with an answer that makes their day easier, makes their job easier, and provides data for other people to share around campus."

Sims laments that while he does get the rare request for printed maps, they are often already outdated while they’re still warm from the printer due to the speed at which data changes at Tech.

Georgia Tech’s Center for Education Integrating Science, Mathematics and Computing (CEISMC) and STEM@GTRI have launched a pilot program for rural Georgia school districts that provides high schoolers with access to interactive modules in the areas of coding, cybersecurity, artificial intelligence, sensors and data visualization. The participating school districts for the 2022-2023 academic year include Cartersville City, Chattooga County, Effingham County, Fayette County, Gordon County, Haralson County, Liberty County, and Walker County.

The initiative, called the Computer Science for Rural Georgia High Schools Pilot, launched in summer 2022 and has been supported with funding from the State of Georgia. Roughly 400 students have participated in the pilot to date and up to 600 are expected to participate in total.  

“This program is increasing the exposure of opportunities in computer science and fields in the areas of science, technology, engineering and mathematics (STEM) for a wider range of students, which is pretty exciting for us,” said STEM@GTRI Director Leigh McCook.

The five modules, which include introduction to coding, advanced coding, principles of cybersecurity, sensors and data visualization, and foundations of artificial intelligence, were developed based on input from the participating school districts and tap into Tech’s areas of expertise. Each module is two weeks in duration and is taught virtually by a Georgia Tech faculty member in collaboration with the classroom teacher, who is in person.

For the introduction to coding class, students learned the basics of coding and then were able to develop and deploy code to the Georgia Tech Robotarium, which is a remotely accessible swarm robotics research platform that is free and open to anyone. The advanced coding class is based on EarSketch, a free educational programming environment developed at Georgia Tech that is designed to teach coding in two widely used languages, Python and JavaScript, through music composing and remixing.

The modules on cybersecurity, artificial intelligence and sensors and data visualization are aimed at addressing current workforce development needs in Georgia. 

With an economic impact of nearly $54 billion, Georgia’s technology sector accounts for 6% of the state’s total workforce, according to recent data from the Computing Technology Industry Association (CompTIA). Georgia’s estimated net tech employment for 2021 was 281,666 workers, a gain of 4,219 net new jobs year-over-year, according to CompTIA’s latest data.  

The number of software, programming, web and quality assurance (QA) occupations led the state for 2021, at 60,863. IT support specialists and repair technicians followed at 25,517; cybersecurity and systems engineers ranked third, at 24,076, per CompTIA.

After the week-long instruction concluded, students completed a project where they solved a real-life problem facing their communities with the technologies they learned about. Then, Tech faculty and students provided the teams with feedback on their projects. 

“It’s a little bit like a Shark Tank environment, where the students receive professional feedback on their projects,” said CEISMC Director Lizanne DeStefano, who also serves as a professor of psychology at Georgia Tech.

Participating teachers said the pilot has given their students the opportunity to make a tangible connection to many valuable computer science topics. 

“Day one was awesome!,” said Stephanie A. Ratliff, a teacher at Chattooga High School. “I just can’t say thank you enough to GTRI and Georgia Tech for allowing us to be a part of this pilot venture.”

Gerald Nelms, a teacher at Bradwell Institute, a high school located in Liberty County, added: “My students were exposed to a wide world of possibilities that exist in computer science. We cannot wait for future collaborative efforts.”  

Once the pilot concludes at the end of the current fiscal year, DeStefano and McCook said they are eager to scale the program to more districts and create a resource repository for participating districts to draw from at any time. 

“At the very minimum, we will develop this into an educational resource and continue to host it on our websites,” DeStefano said. “If there is continued funding, then we would be interested in refining the five modules and offering them to a larger number of districts.”

STEM@GTRI is GTRI's K-12 outreach program. Funded by the State of Georgia, the mission of STEM@GTRI is to inspire and engage Georgia educators and students by providing access to experts in STEM fields. CEISMC is a unit with Tech’s Office of the Provost that serves as the primary connection point between faculty and students and the preK-12 STEM education community.

Writer: Anna Akins 
Designer: Toya Ejike 
Photo Credit: CEISMC 
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $800 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

If they’re successful in using an artificial intelligence technique known as machine learning to associate potentially dozens of factors with the formation of tornadoes, the work could dramatically improve the detection of severe storms – and reduce false alarms.

“This is a great opportunity to apply machine learning to take advantage of the severe storm reports available for the past several years,” said Levi Boggs, a research scientist at the Severe Storms Research Center (SSRC) at the Georgia Tech Research Institute (GTRI). “We can feed all of this information, potentially 30 or 40 different predictors, into the machine learning models and train them to identify patterns that we could potentially use to predict when tornadoes will form. Using AI, we can take on tasks that would be too challenging for humans alone.”

Using data from their ground-based lightning mapping array, the researchers also are studying “jumps” and “dives” in lightning activity to see how they may help predict the formation of tornadoes.

Overcoming the Challenges of Radar

Forecasters now rely on weather radar to identify tornadoes and predict which storms may spin them off. But in areas such as North Georgia, topographical features such as mountains can limit the ability to see lower portions of potentially-dangerous storms, while the time required for radars to update their views can cut into warning times. Electromagnetic interference also can create confusing radar results, and during large severe weather outbreaks stretching across hundreds of miles, there can be multiple storms that must be watched for signs of tornadic activity.

As a result, the development of tornadoes can be missed, while false alarms may lead citizens to disregard warnings – or wait too long to seek shelter. Based on research conducted so far, Boggs believes warnings based on machine learning techniques could be significantly faster and more accurate – and offer the potential to automate the tracking of the storms.

“With radar-based methods, there can be a high false alarm rate, as much as 60 or 70 percent,” he said. “At the same time, the probability of detection can be as low as 50 or 60 percent, which means a lot of tornadoes are missed. With these machine-learning techniques, we expect to improve on both detection and false alarm rates.”

Training Machine Learning with Detailed Storm Reports

So far, researchers have trained their machine learning system on data from 62 tornadoes resulting from 40 different storms in Georgia. In the Peach State, tornadoes commonly pop up from squall lines of storms, though supercells – larger rotating behemoths more often seen in the Midwest – also bring tornadoes into the state.

Supercells can spawn more powerful tornadoes – EF3, EF4, and EF5 – which are more dangerous to humans and destructive to property. But squall line tornadoes can also be deadly, even if they create less powerful EF0, EF1, and EF2 tornadoes, and lines of storms capable of producing them may extend across multiple states.

“One of the main benefits of this machine learning technique is that by using data from the geostationary lightning mapper on the GOES satellite, you would be able to avoid the limitations of radar,” he said. “Using satellite data, you have a huge field of view without the terrain blockages, and you can detect tornadoes over a huge distance – potentially the entire continental United States.”

Using the technique, Boggs and his colleagues are evaluating as many as 40 different parameters to see which ones may be relevant to predicting tornado formation. Among them is the pattern of electrical charge within the storms, which he compares to a genetic profile.

“A typical thunderstorm may have two or three charge regions, but the supercells could have a dozen or more separate regions,” he said. “It’s really complicated to see what’s going on with the lightning because those complex charge structures will create different types of discharges. The flash rate can be just noisy.”

Despite the potential advantages of satellite tornado prediction, Boggs believes forecasters will likely continue to use existing radar techniques, supplementing them with new technology as it develops. GTRI has submitted proposals to funding organizations to continue testing the machine learning tool, which also could be useful to countries that lack the weather radar network available to forecasters in the United States.

Analyzing Lightning ‘Jumps’ and ‘Dives’

Satellite data and machine learning aren’t the only approaches SSRC researchers are using to identify where tornadoes and other severe weather will pop up.

For several years, GTRI has operated the ground-based North Georgia Lightning Mapping Array (NGLMA) that tracks lightning bursts in North Georgia, centered on the Atlanta metropolitan area. Researchers are using radio-frequency emissions recorded by the array to study lightning flashes in an effort to correlate “jumps” – increases in lightning occurrence – and “dives” – reductions in frequency – with the development of severe storms.

The ground-based array – one of several operating in the United States – provides information not available from satellites, so the two sources are complementary, providing both optical and radio-frequency data.

The array was deployed by John Trostel, director of the SSRC, and correlates data on electromagnetic energy produced by the lightning bursts with precise timing and location information. The network of 12 ground stations tracks both lightning that interacts with the ground as well as bursts that stay in the clouds – which account for 75 percent of all lightning – providing a detailed map of electrical charge in the atmosphere.

“What we are looking for is a rapid increase in how many flashes there are over a brief period of time, on the order of a couple of minutes,” said Jessica Losego, an SSRC research meteorologist who is using a NASA-developed algorithm to study the phenomena. “If you see a jump, you can feel somewhat confident that you’re going to soon have some type of severe weather that may include damaging wind, hail, or a tornado. Analyzing this can help with all modes of severe weather, not just tornadoes.”

Losego is among the weather researchers worldwide who are also studying dives, sudden declines in lightning rates, though it’s not yet clear how – and if – they may help forecasters. The dives in lightning activity may serve as yet another indicator of the strength of a storm and how it may be changing.

How Georgia’s Severe Weather Is Different

After a tornado killed a dozen people in North Georgia in 1998, the SSRC was created by the state of Georgia to develop improved means of providing early warning of tornadoes and severe storms. Beyond topographical issues, Georgia’s tornadoes can differ from those of neighboring states in other ways, Losego noted.

“A lot of our storms come through later in the day, which means there’s less sunlight to provide energy to the storms,” she said. “The storms may start in Mississippi early in the day and may fall apart by the time they get there, but they are still dangerous. Storms that arrive late in the day or evening can make it more difficult to warn citizens who may be asleep when tornadoes are detected.”

Data gathered by the NGLMA is shared with National Weather Service (NWS) forecasters in Peachtree City, providing an additional source of information for its forecasts.

“Our goal is to provide another tool that the NWS can use to provide more warning and have more confidence in that warning,” Losego said. “Data from our lightning mapping array goes directly into their systems, and we will share what we learn about using information from jumps and dives that could improve warnings to Georgia citizens.”

The NGLMA now covers North Georgia. Because the southern part of Georgia is out of the range of the NGLMA network and can have a different set of weather conditions, the researchers would like to establish a second array to track severe storms there.

Research Supports SSRC Goals

The SSRC was created through funding from the Georgia Emergency Management Agency (GEMA), the Federal Emergency Management Agency (FEMA), and the state of Georgia to serve as a focal point for severe storm research in Georgia.

“The SSRC serves the state of Georgia by actively developing alternative methods for detecting and forecasting severe local storms and exploring improvements to existing storm prediction and sensor technology,” said Trostel. “We are utilizing the latest in machine learning, data analysis, and other technologies to support the goals of keeping Georgians safe from severe storms.”

Writer: John Toon (john.toon@gtri.gatech.edu)
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $800 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

“When I was in elementary school and was asked, ‘What do you want to be when you're older?’ my answer was a game developer," he said.

In his final semester at Tech before graduating with a B.S. in computer science, Carr won the Best Student Game award at the Independent Game Festival (IGF) in San Francisco for his adventure game, Slider.

He began studying game development in high school, leading him to eventually enter a game jam –– a competition that tasks participants with creating a game from scratch in a short time. Though he didn't find immediate success in his first contest, Carr kept going.

The game, a PC title in which players solve puzzles and rearrange maps to help reconnect humanity, emerged from a competition in November 2021. The game received positive feedback after the jam, and while Carr felt like there was more to be done, he nearly let the project fall by the wayside.

"I remember along the way, there was a lot of doubt in me," he recalled. "I asked myself, ‘Should I carry this through all the way?’ I remember someone telling me that you have to trust yourself at the start of the vision you set out on because while you're working through it, you'll doubt yourself a lot. And, I just did that and kept working on it."

Carr looked to the Tech community for help and pitched the game to the Georgia Tech Video Game Development Club (VGDev) in January 2022. Work on the game continued over the next two semesters, and they submitted Slider to IGF, which receives over 600 entries, later that year. Carr and the team didn't expect a response, but to their surprise, in early January, they were named one of six finalists in the student category.

Over spring break, Carr and six other VGDev members went to San Francisco for the conference. Carr still didn’t believe that winning was a possibility, so when Slider was announced during the award ceremony, he was genuinely shocked. He took the stage and reflected on the hard work that went into the game's development by nearly 30 individuals over the years. 

While winning was a highlight of the trip, Carr found himself similarly enthralled with the universal language of gaming.

"One of the coolest things was seeing how much of an international community there is around game development –– there were all sorts of games from European countries, Latin America, and all over the world. Everyone is making games."

A playable demo of Slider is available on Steam, and Carr plans to leave the link active through development in the hopes of expanding the game's reach. As someone who grew up on PC gaming, he knows the platform is accessible to a vast audience. Despite the recent accolades, he explains that the game is not a finished product, but taking his own advice, he plans to trust his vision and keep working on it.

With one degree in hand, Carr is now interning with Amazon and will return to Tech in the fall to get his master's degree in computational intelligence. He plans to keep game development as a hobby for the time being but admits that he’ll never close the door on pursuing it as a career in the future.

Wenke Lee, John P. Imlay Jr. chair and professor at the School of Cybersecurity and Privacy, will represent Georgia Tech as one of the co-principal investigators for the project, which aims to improve how essential business technologies are protected from evolving cyber threats.

“This NSF AI Institute is a really exciting opportunity because it enables us to explore new ideas and develop novel technical approaches and educational content at the intersection of AI and Cybersecurity,” said Lee. “Our team of researchers in multiple disciplines from several leading universities will be collaborating very closely and I expect to learn a lot from them throughout the project.” 

Lee will collaborate with the second co-principal investigator from Georgia Tech, Xiaoming Huo, A. Russell Chandler III professor in the H. Milton Stewart School of Industrial and Systems Engineering. The two will jointly supervise several Ph.D. students and co-develop new course materials and projects to broaden the impact of their research. 

“Cybersecurity research ascends as a groundbreaking sphere in data science, encompassing a myriad of captivating and rigorous topics,” said Huo. “Pioneering innovation remains crucial, and I am eagerly anticipating the opportunities that this grant will bring, ushering in a new epoch in the field.”  

Researchers will work across disciplines to develop new approaches to artificial intelligence that is informed by and works with security experts. The AI tools developed by ACTION will perform security tasks quickly and accurately while anticipating potential moves made by adversaries. The AI will counteract the possible attacks in a way that protects computer network security and ensures people’s safety.  

The work done by ACTION will also include an outreach component. The results from the project are anticipated to innovate education from K-12 to postdoctoral students. Many of the new AI and cybersecurity tools developed will be applied to workforce development, collaboration opportunities among academic organizations, and industry partners.  

Researchers from the University of California, Santa Barbara will take the lead on this project. They will collaborate with Georgia Tech, Purdue University, the University of California, Berkeley, University of Chicago, University of Washington, University of Virginia, Rutgers University, University of Illinois Urbana-Champaign, University of Illinois Chicago, and Norfolk State University. Georgia Tech will receive $1.5 million from this grant.

The first eight projects in the inaugural cohort, along with the seven projects chosen last year, have been a great success. In this third year, the fellowship is expanding to include an additional seven projects that will further the research collaboration across Georgia Tech’s schools and colleges.

“We really want connectivity to manifest through research collaborations, and it’s advantageous for us to reach into the broad wealth of and depth of talent across the academic schools,” said Mark Whorton, GTRI’s chief technology officer. “From the theoretical research done on campus into the applied research we do at GTRI, we're seeking to take those great capabilities and bring applications into the national security space.”

Across the seven selected fellowship awards for the upcoming academic year, researchers from GTRI labs will co-advise students along with a Georgia Tech faculty member. This year’s projects will lead to innovations in everything from electronic warfare systems, artificial intelligence/machine learning, autonomous systems, and protein sequencing to international policy.

Faculty Research Pairs and Proposals 

What: Reconfigurable Metasurfaces for High-Power Microwave Systems and Emerging EM Spectrum Operation Concepts

Who: Dr. Nima Ghalichechian, Dr. Joshua Kovitz, Walter Disharoon

Unit: School of Electrical and Computer Engineering; Advanced Concepts Laboratory (ACL)

Why It Matters: Reconfigurable metasurfaces have the potential to improve high-power microwave (HPM) systems, enabling applications such as adaptive beamforming and beam shaping, frequency tuning, and polarization timing for use in radar, communication systems, directed energy, and other electronic warfare systems. This research proposes to develop reconfigurable metasurfaces using vanadium dioxide (VO2) switch technologies for HPM systems, and demonstrate a reconfigurable reflectarray (RRA) and high-power limiter metasurface.

“Phase-change materials offer a completely new paradigm for the ubiquitous RF switch, a fundamental building block in sensor and electronic warfare systems,” said Kovitz and Ghalichechian. “As a part of this joint effort, we plan to design, fabricate, and test novel reconfigurable and high-power microwave structures based on these phase-change materials.”

What: Interactive Decision-making and Resilient Planning for Long-Horizon Collaborative Manipulation in Complex Military Environments

Who: Dr. Ye Zhao, Dr. Stephen Balakirsky, Maxwell Asselmeier

Unit: School of Mechanical Engineering; Aerospace Transportation & Advanced Systems Laboratory (ATAS)

Why It Matters: Collaborative manipulation, as a class of general-purpose autonomous systems, provides an expansive set of desirable capabilities to perform complex tasks in highly unstructured environments. These autonomous systems could operate in dangerous environments that are inaccessible to first responders, saving labor and reducing the risk to human life. This will open the opportunity of enabling human operators to focus on high-level, critical decisions.

“This fellowship will support human-robot teaming with a robot that has a high level of autonomy along with a sense of touch,” said Balakirsky. “This combination will allow a human operator to provide tasking of dexterous manipulation tasks to the robot without the burden of teleoperation or constant process monitoring. This system has wide-ranging applications from search and rescue to manufacturing.”

What: Trustworthy Edge Systems for Video Analytics: Robustness, Safety, and Resilience

Who: Dr. Ling Liu, Dr. Margaret Loper, Connor Geurin

Unit: School of Computer Science; Information and Communications Laboratory (ICL)

Why It Matters: Video as an edge Artificial Intelligence (AI) service will be a crucial component in many cyber-physical systems and applications. However, most of the video analytics today are typically done in the Cloud, which incurs overwhelming demand for bandwidth. This research is centered on developing trustworthy edge systems for video analytics, including developing the theory, algorithms, and techniques for boosting the robustness of real-time object detection. This will ensure safety and resilience against different types of disruptions and compromises.

“The proliferation of mobile computing and Internet of Things has created a paradigm that pushes computing tasks and services from the network core to the network edge,” said Loper. “Pushing AI to the edge is seen as a promising solution for processing the massive amounts of small data generated by these devices. The findings of this research could fundamentally change how AI-enhanced edge systems will be designed, developed, and deployed, and could lead to a new generation of security and safety-enhanced edge systems.”

What: Model-based Reinforcement Learning for Policy-perspective Explainable and Trusted Artificial Intelligence

Who: Dr. Sehoon Ha, Dr. Robert Wright, Morgan Byrd

Units: School of Interactive Computing; Cybersecurity, Information Protection, and Hardware Evaluation Research Laboratory (CIPHER)

Why It Matters: The emergence of capable artificial intelligence (AI) that can make sequential strategic decisions via deep reinforcement learning (deep RL) has revolutionized various fields, including computer games and robotic control, but they have not yet impacted safety-critical domains such as power grid control, medical treatment, and autonomous driving and far from real-world deployment. This research investigates scalable model-based RL approaches for explainable and trusted AI to develop explainable AI learning frameworks that can be applied to these safety-critical domains.

“AI technologies are becoming more and more capable every day and are on the verge of revolutionizing many fields and industries,” said Wright. “However, AI models are prone to mistakes, and their reasoning can be very opaque, leading to a [reasonable] lack of trust. This effort investigates novel explainable AI approaches for Reinforcement Learning (RL) to improve trust and practicality. Our intent is to develop model-based RL algorithms that can explicitly describe why it is making its decisions, visualize or describe what it expects to happen, and provide counterfactual examples for why it chose not to make decisions.”

What: Two-dimensional Nanopore Sensors for Real-time, Single Molecule Protein Sequencing

Who: Dr. Eric Vogel, Dr. Katherine Young, Noah Baughman

Units: School of Materials Science and Engineering; Cybersecurity, Information Protection, and Hardware Evaluation Research Laboratory (CIPHER)

Why It Matters: There is a significant need to develop rapid protein sequencing technologies that can be used by the warfighter in the field to identify the impact of biological warfare agents or to provide physiological monitoring to enhance soldier performance. A technology to rapidly sequence the primary and secondary structure of proteins at the single-molecule level in real-time does not currently exist. The objective of this work is to develop a rapid protein sequencing prototype technology based on two-dimensional (e.g., graphene, MoS2) nanopore sensors that can be used by the warfighter in the field and enable future research programs which apply this prototype to perform full protein sequencing.

“There is a significant need to develop rapid protein sequencing technologies that can be used to identify the impact of biological warfare agents or to provide physiological monitoring to enhance human performance,” said Vogel and Young. “This fellowship will support the fundamental research necessary to develop nanopore electrochemical sensors based on two-dimensional materials to rapidly sequence the primary and secondary structure of proteins at the single-molecule level in real-time.”

What: Generating Geopolitics: AI, Disinformation, and the Future of National Security

Who: Dr. Jon Lindsay, Mr. Nicholas Nelson, Dennis Murphy

Units: School of Cybersecurity and Privacy, Sam Nunn School of International Affairs, and School of Public Policy; Electronics, Optics, Systems Directorate (EOSD)

Why It Matters: The use of Artificial Intelligence/Machine Learning (AI/ML) in national security has the potential to enhance our ability to protect national interests greatly. However, there are also potential challenges and risks associated with this technology, such as the potential for bias or misuse. This research will engage in a multidisciplinary study that will bridge the gap between disparate research fields and reintroduce relevant security-related concepts from the social sciences. This will result in the generation of scientifically-grounded potential use cases for the technology in the support and protection of national interests.

“As AI/ML capabilities and use cases continue to evolve, it is critical for defense and national security actors to better innovate, scale, deploy, and integrate AI and autonomy-based technologies to form agile, system-wide solutions,” Nelson and Lindsay said.

What: Unmasking the "Status dilemma/competition" of the triad powers (Russia, China, and United States) in offensive-defensive behavior

Who: Dr. Adam Stulberg, Dr. Theresa Kessler, Megan Litz

Units: Sam Nunn School of International Affairs; Advanced Concepts Laboratory (ACL)

Why it matters: Unveiling the misperceptions of offensive and defensive signaling is needed in a time when offensive and defensive capabilities are becoming ever more difficult to decipher as technology is evolving. The goal of this research is to shed light on how misinterpreting states’ status can lead to international conflict and expand the initial scholarship that is starting to gain traction within the political science and security studies communities. Understanding and attempting to codify intention would be of great interest to U.S. strategists and tactical planners and aid in answering vital questions of National Security regarding the status of triad powers. Information of this nature will benefit U.S. leadership, departments, and inter-agencies that navigate relations with Russia and China.

“This fellowship will support the codification of offensive and defensive signals between Russian, Chinese, and American powers using an open-source literature repository,” said Kessler. “This will help unveil misperceptions and decipher intention.”

Writers: Georgia Parmelee, Tess Malone (Georgia Tech Research); Charles Domercant, Anna Akins (GTRI)
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,900 employees, supporting eight laboratories in over 20 locations around the country and performing more than $800 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

The Georgia Tech Research Institute (GTRI) is working to solve this challenge by integrating its Real-time Intelligent Fusion Service (RIFS) into its Forklift Assist System (FAS) for warehouse operations to streamline efficiencies and enhance worker safety. RIFS, which is a part of FAS, was built with the cross-platform game engine Unity and produces spatial information about a room and then displays that information as meshes on a device, such as a desktop computer or tablet. FAS also includes a camera system that has additional forklift assistance features.

This project has been supported by the U.S. Marine Corps and U.S. Navy, and has also been tested at the Marine Corps Logistics Base in Albany, Georgia.    

According to recent estimates, the average U.S. warehouse wastes 6.9 weeks a year on unnecessary motion, which costs the industry $4.3 billion, or 265 million hours of labor annually. Additionally, in 2020, the latest year for which statistics are available, there were 5.5 injury and illness cases per 100 full-time workers and 21 fatalities in the warehouse and storage industry.  

RIFS would be incorporated into a warehouse's order management system and provide real-time information about everything going on in the warehouse. For example, a forklift operator could display RIFS on their tablet device and it would help them navigate to their pick up and drop off locations while ensuring they steer clear of obstacles.

"The system would know where all the other forklifts and people are in the warehouse and have route planning functionality," said Stephen Balakirsky, a GTRI principal research scientist who is leading the project. "A lot of warehouses have one-way aisles and it can be difficult for humans to determine the most efficient path to take, but RIFS could automatically determine that for you."

RIFS works by creating a grid of a particular environment with individual grid spaces that indicate which areas are traversable or not, explained GTRI Research Scientist Emily Strube, who has expertise with RIFS. The software then utilizes a pathfinding algorithm to determine the effort or "cost" required to move from one grid space to another and maps out the most efficient path possible. People and objects are shown as meshes, or geometric objects, in RIFS.  

For example, in a warehouse, the algorithm would determine how much effort is required for a worker to move from their current position to their pick up or drop off point and outlines the most expedient route, Strube said.

She added that the algorithm can be adjusted to accommodate changes in a warehouse's workflow and forklift routes.   

FAS' camera system would help forklift drivers move towards a pallet and secure it without damaging the warehouse's infrastructure or other pallets and items. The camera system would also have a backup feature to give drivers additional awareness as they move throughout the space in reverse.  

To reduce the risk of forklift loads colliding with the top of doorframes, GTRI is also considering developing and installing sensors near doorways to alert drivers of imminent collisions.  

"There are lots of different safety features that could be added to this project," Balakirsky said. 

Additionally, GTRI seeks to further utilize RIFS to provide remote inspection and validation of inventory through virtual reality (VR) technology, where the warehouse maps and images would be created by autonomous robots.

RIFS has been incorporated into several other projects, including GTRI's Independent Research and Development (IRAD) of the Year winner for fiscal year 2022, which Strube leads. That project seeks to increase the situational awareness of troops on the ground.

Writer: Anna Akins 
Photo Credit: Stephen Balakirsky, Emily Strube 
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

ICARIS is motivated by the fact that, while the scope of the cyber threat against critical infrastructure is understood, capabilities to address the threat are spread across numerous organizations, involving diverse academic, government, and industry stakeholders. Moreover, there is a critical shortage of skilled personnel with an understanding of both cybersecurity and the underlying physical systems.

"Georgia Tech is proud to partner with PNNL to accelerate vital infrastructure research that will benefit people nationwide," said Chaouki Abdallah, executive vice president for Research at Georgia Tech. "This collaboration leverages Georgia Tech’s and PNNL’s complementary strengths, catalyzes new and exciting research directions, and will serve as a national resource for all infrastructure sectors."

According to David Manz, PNNL’s co-director for the institute, and Georgia Tech Research Institute’s (GTRI) Alexa Harter, the primary goals of ICARIS are to: (1) perform translational R&D that moves innovative concepts towards implementation into operational environments; (2) develop the future workforce; and (3) provide advice and solutions to communities, states, federal agencies, and businesses. The institute will guide technology development roadmaps for critical infrastructure sectors while influencing a whole-of-government approach to cybersecurity for these sectors, directly supporting national strategies such as the White House’s Industrial Control Systems Cybersecurity Initiative.

“This collaboration brings together the capabilities of PNNL and Georgia Tech to accelerate our combined contributions to the security of the critical infrastructure community,” said Deb Gracio, PNNL’s Associate Laboratory Director for National Security.

“The partnership will offer a distinctive value proposition founded on three areas of expertise that Georgia Tech and PNNL can uniquely combine: threat intelligence and assessments, deep technical understanding of cybersecurity and engineering domains, such as electric power systems, and an ability to translate threat-informed technologies and technical knowledge into operational impact,” according to GTRI Director James Hudgens. 

The collaboration will cement the growing relationship between Georgia Tech and PNNL in cybersecurity and advanced computing, establishing the foundation for long-term research partnerships and major new programs. 

"Ensuring the cybersecurity of the nation’s critical infrastructure is of paramount importance if we are to continue to prosper as the result of our investments in technology. This joint institute will leverage Georgia Tech’s 20-year-long commitment to cybersecurity to further develop innovative research directions and educational curricula aimed at improving the security and privacy of systems, such as the electric power grid, which are so essential to our everyday lives," added Michael Bailey, Chair of Georgia Tech's newly formed School of Cybersecurity and Privacy.

“This partnership is an important and exciting step forward in building multidisciplinary teams of researchers who are dedicated to protecting our nation’s energy infrastructure so our future energy systems are resilient, reliable, and secure,” said Jud Virden, PNNL’s Associate Laboratory Director for Energy and Environment.

The joint institute will build on PNNL’s strengths in advanced computing and data science for security, grid controls and cyber defenses, and vulnerability assessment for critical infrastructure. The institute will also provide a pathway to apply PNNL’s significant capabilities in resilient controls for the power grid, critical infrastructure test ranges, and AI methods to automate defensive maneuvering and threat discovery. 

“This partnership falls directly out of the Georgia Tech Research Next Strategic Plan, which recognizes that we must increasingly partner with a broader spectrum of stakeholders to truly deliver the positive impact we have committed to on our planet’s toughest problems,” said SEI Executive Director and Research Next co-chair Tim Lieuwen. “We are committed to putting in the work to be a great partner.” 

Georgia Tech has initiated a national search for the joint institute co-director. Application information can be found here. Lee Lerner, a GTRI principal research engineer, has been named interim co-director. 

GTRI Communications

Georgia Tech Research Institute

Atlanta, Georgia USA

About GTRI: The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, the state, and industry. For more information, please visit www.gtri.gatech.edu.

GTRI recently teamed up with the University of North Georgia, the Georgia Cyber Center and the Georgia Department of Education to promote CyberStart within the state.

For the latest cycle of CyberStart that ran from October 2021 to April 2022, Georgia led the nation in participants, with 6,383 students from 274 Georgia high schools competing. That represents a whopping 564% increase in Georgia student participation compared to the previous cycle. A total of 45,962 students competed nationwide.

CyberStart is an online game designed to help students learn about cyber topics as they complete fun puzzles and challenges. The game allows students to take on the role of cyber protection agents as they solve cybersecurity-related puzzles and gain exposure to code breaking, programming, networking and digital forensics.

Students can play the games at their own pace and all students in grades 9-12 are able to participate.  

In addition to building their cyber skills, students also have the opportunity to compete for scholarships and cash prizes.

The CyberStart America in Georgia taskforce gave out $102,497 in cash prizes to top performing students, schools and districts during the 2021-2022 competition season.

"There's a real need for talented cybersecurity professionals in Georgia," said Tyler Kinner, a GTRI research scientist who is a part of the CyberStart America in Georgia taskforce. "Not only will this program help address the future cyber workforce shortage, but it will also help students become savvier in terms of how they interact with technology."   

There are over 75 cybersecurity companies in Georgia that generate an estimated $2.6 billion annually, according to recent estimates. Georgia ranked in the top seven states for cybersecurity growth potential by Business Facilities Magazine in 2021.

Additionally, there are 700,000 cyber job openings nationally, with about 25,000 of those in Georgia.

GTRI led the development of a standards matrix to communicate the alignment between CyberStart and the Georgia Department of Education's curriculum requirements for Georgia high school students. Since its first use during the 2021-2022 school year, other states’ CyberStart America working groups have followed in developing their own standards-based matrix for this year’s competition season.

"We went through CyberStart to document what knowledge and skills students were learning, and used this to develop a matrix where we demonstrated the alignment between CyberStart and the course standards for Georgia’s cybersecurity pathway courses,” Kinner explained. "That way, we could communicate to teachers and leaders, 'Here's this acclaimed international platform, and here’s how it aligns to our state's standards.'"

Kinner noted that the gamified format of CyberStart has helped spur record participation in the program, and he is optimistic that the Georgia participation numbers for the 2022-2023 academic year will be even higher than last year.

"I think CyberStart has hit a sweet spot in terms of figuring out how to engage high school students with a format that is both educational and entertaining," Kinner said.

Writer: Anna Akins 
Photo Credit: University of North Georgia 
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

To overcome these challenges, Georgia Tech researchers form the core of an interdisciplinary, interorganizational team which seeks to prevent disease outbreaks by integrating the study of human behavior with computational data-driven models. 

Calling themselves BEHIVE (BEHavioral Interaction and Viral Evolution), the group recently received a $1 million National Science Foundation (NSF) grant toward multidisciplinary team formation and novel outbreak prevention research.

“Our goal is to bring together all these terrific researchers from different disciplines to help bring a paradigm shift in the science of pandemic prediction and prevention,” said B. Aditya Prakash, associate professor with Georgia Tech’s School of Computational Science and Engineering (CSE). 

“While epidemic forecasting is compared to weather forecasting, there is an important difference. Unlike weather, our actions and behavior can change the course of an epidemic.”

Prakash is the principal investigator of the $1 million NSF grant. Fellow BEHIVE members include:

• Pinar Keskinocak, William W. George Chair and Professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech
• Thomas Kingsley, Assistant Professor of Medicine and Biomedical Informatics at Mayo Clinic
• Shinobu Kitayama, Robert B. Zajonc Collegiate Professor of Psychology at the University of Michigan
• Ramesh Raskar, Associate Professor at the Massachusetts Institute of Technology Media Lab
• Liliana Salvador, Assistant Professor at the University of Georgia’s Department of Infectious Diseases
• Joshua Weitz, Professor and Tom and Marie Patton Chair in the School of Biological Sciences and Co-Director of the Interdisciplinary Ph.D. in Quantitative Biosciences (QBioS) at Georgia Tech

Prakash emphasized BEHIVE’s primary goal to use its interdisciplinary organization to bridge research methodologies between hard and soft sciences. 

He explained that human behavior was underutilized in epidemic science before Covid-19, largely due to data scarcity and underdeveloped computational technologies. Behavioral dynamics encountered during the pandemic, such as social distancing, mask wearing, and vaccine hesitancy, has provided new research and data that now can be considered in models and simulations.

Here, BEHIVE will develop high fidelity computational models by designing new artificial intelligence and machine learning techniques that bridge human behavior knowledge and traditional epidemiological theory and models.

“It is still an open question of how we can best incorporate human behavior knowledge into the study of pandemics. That is the challenge,” Prakash said. “Our main idea is to better integrate knowledge from psychology and the humanities into pandemic science using novel computational methods.”

BEHIVE originated when team members met through various workshops held in 2020 and 2021. Prakash was an invited organizer of the National Symposium on Predicting Emergence of Virulent Entities by Novel Technologies (PREVENT). 

PREVENT reported that interdisciplinary collaboration was an obstacle in predicting and preventing pandemics. For example, some vocabularies often don’t mean the same thing across disciplines, so a consistent methodology to establish a common language must be developed.

BEHIVE is custom built to solve these challenges PREVENT revealed. Along with a wealth of knowledge learned through past epidemics, each BEHIVE researcher brings to the group experience working across interdisciplinary lines. 

Among the Georgia Tech researchers alone, Keskinocak interfaced with policymakers and the public on measures to slow Covid-19 spread. 

Prakash’s lab led several high-profile Covid-19 forecasting initiatives, including collaboration with the Center for Disease Control and Prevention (CDC).

Weitz teamed with fellow Georgia Tech researchers with the College of Science, College of Computing, and the Wallace H. Coulter Department of Biomedical Engineering to create a predoctoral training program that integrates computational modeling and data analytics into bioscience.

Keskinocak, Prakash, and Weitz together are also faculty in the Institute for Data Engineering and Science (IDEaS), one of Georgia Tech’s ten interdisciplinary research institutes. IDEaS connects research centers and efforts in foundational areas such as machine learning, high-performance computing, and algorithms.

BEHIVE’s $1 million grant is funded through NSF’s Predictive Intelligence for Pandemic Prevention (PIPP) initiative. This program supports high-risk, high-payoff convergent research that aims to identify, model, predict, track, and mitigate the effects of future pandemics.

According to Prakash, the PREVENT symposium’s summary report helped lay the foundation for the PIPP program.

PIPP is a two-phased initiative in which NSF selects to fund 25 to 30 project teams, including BEHIVE, for eighteen months through phase one. However, this does not necessarily limit PIPP’s influence to chosen project teams within academia.

BEHIVE intends to partner with industry, governmental, and non-profit organizations to expand its interdisciplinary, interorganizational network. 

BEHIVE’s nucleus of Georgia Tech researchers connects the group with the CDC, Georgia Department of Public Health, and numerous hospitals across the state. BEHIVE’s other researchers also serve in leading roles at non-profits, such as the Pathcheck Foundation, and top hospitals like the Mayo Clinic.

Along with developing interdisciplinary methodologies, new disease prevention models, and partnering with external organizations, BEHIVE hopes to develop educational training programs. This would ensure their effort last generations to bring about the necessary paradigm change to prevent future pandemics.

“Our initial projects and research the next eighteen months will help us get a sense of research gaps and enlarge our perspective” Prakash said. “We’re approaching PIPP as a science, and we want to lay the foundation of the science by bringing in many people from different fields for the future.”

The cohort consists of 11 officers – two from the U.S. Navy and nine from the U.S. Air Force – who are working in six of the eight GTRI labs. That brings the participation total in the program up to 17 since its inception in fall 2020. The two Navy officers bring fleet experience from operational tours and all nine of the Air Force officers are recent graduates of the U.S. Air Force Academy, who are on track to pursue careers in pilot training, cyber operations, developmental engineering, operations research, and military meteorology.

MGRP Chair Mario Mifsud, who serves as the associate lab director of GTRI's Electro-Optical Systems Laboratory (EOSL), called the program a win-win for all the involved parties, allowing service members to solve applied engineering problems on sponsored, real-world DoD-related projects while providing GTRI with top talent.

"This fall, GTRI is getting seasoned Navy officers who bring a wealth of military experience and Air Force Academy graduates who are really sharp students," Mifsud said. "For the participants, they have the opportunity to earn a paid-for master's degree from a renowned research university on top of gaining real-world experience. Everyone benefits from this program."

MGRP funds its graduate degree program through Georgia Tech's Graduate Student Tuition Remission Plan (GSTRP). Throughout the program, each participant serves as a military graduate research assistant (MGRA), which is the equivalent of a graduate research assistant or graduate teaching assistant (GRA/GTA). The fall 2022 semester is the first time GTRI will cover associated degree fees. Some tuition and fees are also waived due to the MGRA's military status. Book expenses are the MGRA's responsibility.

The MGRP selection process has three components.

First, the military officer must apply to a service sponsoring program, and the program must put the individual on active duty, Permanent Change of Station (PCS) orders to the Atlanta area to participate in MGRP. At the same time, candidates apply to the Georgia Tech STEM graduate degree program of their choice and must be accepted into their desired program to be eligible to participate. Thirdly, candidates must fill out an MGRP program application, which is available on GTRI's MGRP webpage.

Once these three steps are completed, the candidate’s application package is forwarded to the GTRI labs, divisions, and branches that best align with the candidate’s graduate degree program and research area(s) of interest.

Mifsud said that the lab placements represent a best fit and many participants receive more than one offer from within the different GTRI labs. 

In addition to spreading the word about GTRI's science and engineering expertise, Mifsud said MGRP builds a lifelong bond with service members and further strengthens GTRI's relationship with the military – for whom much of its work is dedicated.

"People say there is no free lunch," Mifsud said. "But in this program, there is. All of the players, all of the stakeholders, get something more economically than they would if they were doing things on their own."

Writer: Anna Akins
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

The five-week program, which was held June 13 to July 22, hosted 65 high school students from 13 Georgia school districts who were selected from an application pool of 487. The students worked under the direction of 34 professionals at the Georgia Tech Research Institute (GTRI), who represented seven of the eight GTRI labs. At the conclusion of the program, all students presented the results of their work in a daylong event for GTRI leadership, mentors, and special guests.

The ultimate goal of the program is to provide students with real-world experience in the fields of science, technology, engineering, and math (STEM) and to create awareness of future STEM career opportunities. 

"It's an incredibly rewarding experience to work with such talented students and see what they are able to accomplish in five weeks," said Therese Boston, a GTRI senior research associate and co-director of the program. "I can't wait to see how the interns build upon the skills they have learned during their internships in school and in their future careers." 

The participating students represented the following Georgia school districts: Atlanta Public Schools, Cobb, DeKalb, Dougherty, Douglas, Fayette, Forsyth, Fulton, Gwinnett, Harris, Henry, Houston, and Marietta City Schools.

Though many participants previously had some experience with STEM-related topics prior to the internship experience, such as coding and robotics, they said the program further strengthened their skill sets in those areas by requiring them to apply that knowledge to tackle real-world challenges designed by GTRI employees who served as mentors.   

For one project, students designed and evaluated modifications for a suction-based robotic end effector, a claw-like device attached to the end of a robot's arm that interacts with the environment and is capable of picking and placing raw chicken products in a poultry processing facility.

The project required students to design experiments and utilize computer-aided design (CAD) software, including SolidWorks, to create components on a 3D printer. 

One participant, Kari Britton, who is a student in the Fulton County school system, said gaining exposure to SolidWorks while also observing the real-world applicability of her research was invaluable.

Another group developed an augmented reality (AR) facial detection program used to accurately detect and register users into a database. The program would be intended primarily for the education field, helping professors keep track of their hectic classrooms - such as recording when students check in and out of class. The students also incorporated a hand-detection feature into the program to detect when a student raises their hand and quickly notify the professor.

One student in the group, Bhoomi Kotharkar, a student in the Forsyth County school system, said this project not only helped her become more comfortable working with emerging technologies, but it also showed her what a STEM career could look like.  

"One of the most beneficial aspects of the internship was getting hands-on experience with AR, cybersecurity, and cloud computing," Kotharkar said. "It also showed us what our future could look like if we choose to pursue a STEM career."

Robert Clark, a GTRI senior research scientist who served as a program mentor, said in addition to teaching students STEM-focused concepts, the internship also reinforces the importance of teamwork, thinking critically, and asking the right questions. 

"The program offers a number of features that reflect a real workplace," Clark said. "I think the best thing about STEM@GTRI is that it gives students a chance to learn what it’s like to go into a problem in depth, experiencing both the challenge and the satisfaction of really wrapping your mind around technical content."

STEM@GTRI is GTRI's K-12 outreach program. Funded by the State of Georgia, the mission of STEM@GTRI is to inspire and engage Georgia educators and students by providing access to experts in STEM fields.

Writer: Anna Akins
Photos: Christopher Moore
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

AIS uses signals from transponders operating on the ships to help their captains avoid collisions when the vessels are outside of busy ports. Because AIS is based on an open standard developed many years ago, the U.S. Navy's Battlespace Awareness & Information Operations Program Office (PMW 120) realized the system needed hardening to help address current cybersecurity conditions and expectations.

GTRI researchers were initially asked to evaluate potential vulnerabilities of the system, and then to develop an add-on software system, called Bifrost, which works with AIS to filter messages from ships, guard against potentially malicious messaging, and provide critical alerts to ship captains. The Bifrost system has been delivered to the Navy’s Battlespace Awareness & Information Operations Program Office, and is now undergoing evaluation – a step on the way to potential deployment.

“The goal of AIS is to avoid collisions, and everyone works together to contribute information about where their ship is and which way they are headed to make sure everyone can predict where they will be,” explained Shelby Allen, a GTRI research scientist who led the project. “Being able to trust the information being provided is important to ensuring the safety of maritime traffic worldwide. Along with GPS, AIS plays an integral role in how our forces operate across the seas.”

Information for AIS comes from transponders on each ship that provide such information as the GPS-based location coordinates, heading, and speed. The transponders use a common and open protocol, but equipment errors and other factors can affect the accuracy of what’s reported. Bifrost helps filter transponder information coming into Navy ships.

“The goal of the application we developed was not to get in the way of the existing system, since it is a critical path for downstream systems,” Allen explained. “We wanted to look for both accidental issues with the incoming transmissions and the potential for deliberate misuse.”

Because of the critical nature of the communications, the Bifrost system was designed to extract useful information from ship transmissions even if they don’t necessarily meet all the specifications of the protocol.

“The majority of what we see that looks like a transmission not abiding by the specifications are accidental formatting issues,” he said. “Filtering this information needs to have a certain amount of tolerance for what can go wrong and still have the messages provide useful information.”

Bifrost can detect deliberate misinformation, such as location updates that suggest speeds impossible for vessels to attain. “Ships can only accelerate and decelerate at certain rates, and there are some examples of egregious misuse of location information,” Allen said. “One of our goals was to detect messages less likely to be real GPS-based messages.”

Beyond cybersecurity hardening, Bifrost enhances how the system handles emergency alerts, which may not receive sufficient visibility in the original AIS interface.

“There are safety-related messages that by protocol should be addressed immediately,” Allen said. “We worked to make sure that these alerts had the smallest chance of being an annoyance. When someone did need to review the alerts and needed additional information, we made it as easy as possible to do.”

Because Bifrost was intended to be a working software system with an important safety mission, PMW 120 requested the researchers to carry development further than often happens with research projects. “We had to make sure that this was something that could quietly and reliably run for a long time in a performance environment,” he explained.

That reliability and operational testing extended a bit further than Allen originally expected – to ten days on a U.S. Navy guided-missile destroyer off the coast of California – and to bunk space reserved for researchers from organizations working on projects that required real-world testing at sea.

“At first, everybody was kind of learning their way around the ship and making sure they weren’t in anyone else’s way,” Allen said. “We slept in standard quarters and ate the same food everybody else onboard did. We were in the thick of operations on a Navy ship.”

Below deck, where Bifrost was operating, Allen and GTRI colleague David Myers at first lost track of time.

“One of the things I didn’t anticipate ahead of time was how optional it was to be outside,” Allen said. “When I imagined being on a ship, I imagined a huge deck with people there all the time. That was not true at all. The vast number of people are working beneath the deck, and there are multiple levels. There was a point I realized that it had been 24 hours since I had seen the sun.”

The researchers scheduled times to work with the operators of the system, knew when mealtimes were, participated in safety-related exercises, and took advantage of workout facilities – which required some adaptation to the rolling of the ship in the waves.

“There were beautiful, starry nights with absolutely no light pollution,” he said. “Dolphins were following the ship, just like in documentaries. It was a once-in-a-lifetime experience.”

Writer: John Toon (John.Toon@gtri.gatech.edu)

GTRI Communications

Georgia Tech Research Institute

Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

GTRI's Cybersecurity, Information Protection, and Hardware Evaluation Research (CIPHER) Laboratory has participated in capture the flag (CTF) and hackathon events since spring 2021, winning monetary prizes and prestige in the process.

In March 2021, GTRI won $10,000 and placed 2nd in the U.S. Navy's HACKtheMACHINE event, where participants attempted to hack commercial maritime electronics intended for laboratory use to test their vulnerabilities. In December 2021, GTRI was a top 4% finisher in the U.S. Air Force and U.S. Space Force's Hack-a-Sat 2 event, where participants learned how to reduce vulnerabilities in space systems and make them more secure. GTRI in May 2022 placed in the top 5% of the final round of the Air Force and Space Force's Hack-a-Sat 3 competition.

Though the terms “hackathons” and “CTFs” are often used interchangeably, CTFs refer to team-based competitions in which participants use cybersecurity tools and techniques to find hidden clues or flags. The team that finds the most clues or flags, which are hidden in purposefully-vulnerable programs or websites, during the event wins. Hackathons, meanwhile, are events in which developers, designers, and even non-technical people collaborate to build new programs and technologies and do not necessarily involve vulnerability discovery. Most hackathons and CTFs are open to all students, researchers, and professionals across the world.

Chris Roberts, a GTRI principal research engineer who leads CIPHER's Embedded Cyber Techniques branch, said CTFs allow students and faculty of all skill levels at Georgia Tech and GTRI to work together to address issues impacting the cybersecurity field.  

"CTFs involve challenges that represent real-world issues," Roberts said. "What I really like about them is they give seasoned engineers the ability to impart knowledge on more junior-level engineers. Both groups can work together and learn from each other." 

Many CTF events require participants to figure out how to secure legacy technology systems against sophisticated cyber threats.

Pointing to the example of satellites, which are central to the Hack-a-Sat contests, Roberts said many of these systems are prime targets for cyberattacks because they often use obsolete equipment and may not receive regular security updates. The importance of securing satellites holds relevance for advancing national security. Satellites are especially crucial for military operations in the U.S. and across the world, providing geolocation and navigation services, target detection, missile warning and adversary activity tracking.

"Satellites are high-tech pieces of equipment, but a lot of them were launched decades ago when cybersecurity wasn't as much of a concern," Roberts said.  

In addition to exposing participants to relevant cybersecurity issues, Roberts said hackathons and CTFs reinforce the importance of teamwork and problem solving that extend into the workplace.    

"These events require participants to figure out how to approach a problem, break it down into bite-sized chunks, and test their theories," Roberts said. "When I hire a full-time research engineer at GTRI, I'm looking for their ability to problem solve. I can teach them the technical side of things, but problem solving is much more difficult to learn."

Randi Thorson, a GTRI research engineer who earned an M.S. in cybersecurity from Georgia Tech in 2022, said she most enjoys the "rush of" finding flags during competitions and thinking outside the box when testing systems for vulnerabilities. Thorson has participated in CTFs and hackathons at Tech and GTRI for one year.

"I think CTFs are important because they teach you to look for vulnerabilities," Thorson said. "So, when you're designing a product, you know not only some of the mitigations that need to be put in place to design a secure system, but they also teach you the out of box thinking that an adversary will use to exploit the product."

Similarly, Kennon Bittick, a GTRI research scientist, said CTFs and hackathons help people who are new to computer security ease into the field by solving unique problems and just having fun.

Bittick is a Georgia Tech double alum who earned his undergraduate degree in computer science in 2015 and a graduate degree in computer science in 2018. He has participated in hackathons and CTFs at Tech and GTRI since his freshman year of college. 

"The thing I like most is getting a challenge for a system I have never heard of before and being able to quickly do a deep dive, learn about the system, and solve the problem," Bittick said. "To me, it evokes the classic hacker ethos of quickly learning something cool and making something work."

Writer: Anna Akins
Photos: Ethan Trewhitt
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees, supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

The study was conducted collaboratively by researchers at the Georgia Institute of Technology, the University of North Carolina (UNC), and North Carolina State University (NCSU) and the findings published in the research journal JAMA Network Open. The study methods were based on a mathematical simulation originally developed at Georgia Tech. 

The study evaluated how many Covid-19 cases could be avoided in the Tar Heel State if more people get vaccinated and continue to follow mask and physical distancing guidelines. As of June 3, North Carolina has had 1 million reported cases of Covid-19 and more than 13,000 recorded deaths.

“The main takeaway from the paper is that while the increasing vaccine coverage in the U.S. has a positive impact, we are not really there yet. We still need to follow preventive measures such as mask wearing,” said contributing author Pinar Keskinocak, the William W. George Chair and Professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech and co-founder and director of the Center for Health and Humanitarian Systems.

The caution is well founded when the researchers account for viral mutations, including the variant currently dominant in the United States that was initially identified in the UK and was associated with the surge in Michigan. There, as recently as May 2, the state averaged nearly 3,500 cases a day, according to a June 2 story in Bridge Michigan.

According to one scenario from the simulation, which was populated with data from the state of North Carolina, if 75% of the population gets fully vaccinated but continues to wear masks and socially distance,  there is a sustained decline down to very few new Covid-19 cases over a six-month period.  But, if only 25% of the population gets fully vaccinated and does not adhere to these non-pharmaceutical interventions (NPIs), there could be a steady increase in daily Covid-19 cases, peaking around 8,000 before there is another decline.

Keskinocak points to the Georgia Covid-19 Vaccination Dashboard that tracks county-level differences in vaccination rates based on race as further evidence of the need for caution.

“Our dashboard shows that there has been high variability in the level of vaccination in different geographic regions and communities,” she said. “So even if we say over half of U.S. adults are vaccinated, it’s not uniform across the entire country. This further raises concerns about quickly lifting the NPIs.”

Julie Swann, department head, North Carolina State University’s Fitts Department of Industrial and Systems Engineering, and long-time research collaborator with Keskinocak, concurs. “Current variants are more infectious, and there are still locations with less than 30% of the population vaccinated.”

Swann adds that ongoing spread “endangers people now. It also increases the chance of a future mutation that could increase the risk even to people who are vaccinated.” 
Swann and Keskinocak are two of three researchers who co-founded Georgia Tech’s Center for Health and Humanitarian Systems in 2007, the year they developed a comprehensive, agent-based simulation model for pandemic flu, which has since been at the core of their modeling efforts for Covid-19. Over the years the two industrial engineers have collaborated closely to advance the model and make results available to decision makers as new pandemics emerged.   

“Julie and I have been working on infectious disease modeling for over a decade now. We had developed this agent-based model for pandemic flu and then when Covid-19 hit, Georgia Tech adapted that model to Covid-19 and shared it with Julie’s team at NC State who then modified the model to test additional scenarios and calibrated it with North Carolina data,” Keskinocak said.

“I am grateful that Pinar and I had spent such a long time trying to understand pandemics and improving the science behind them,” Swann added. “If we had not invested that time, we would not be able to have such a fast turnaround and the high participation level that we have this year.”

According to Keskinocak, the simulation model is extremely detailed – in essence, it recreated demographics of the population down to the state’s household size, and even individuals’ workflow, to give a clear picture of how people move from one geographic area to another. 

 “The model is flexible and can change based on the research question being asked, including the current research question, ‘What is the impact of lifting NPIs in increasing vaccine coverage?’”

Swann credited the close partnership with public health partners in Georgia and North Carolina, as well as their work under a grant funded by the Centers for Disease Control and Prevention and the Council for State and Territorial Epidemiologists, with the speed of developing models to test interventions during the pandemic. 

“It’s really key because sometimes we get additional ideas or research questions from the stakeholders with whom we interact,” Swann said.

Both researchers emphasized the critical role students played in advancing these models under tight deadlines while juggling coursework. Between them they estimate more than two dozen graduate students across the partnering institutions have been engaged in the Covid-19 modeling work since the pandemic began. The two also were integral in establishing a professional education HHSCM certificate program.

 # # #

The Georgia Institute of Technology, or Georgia Tech, is a top 10 public research university developing leaders who advance technology and improve the human condition.
The Institute offers business, computing, design, engineering, liberal arts, and sciences degrees. Its nearly 40,000 students, representing 50 states and 149 countries, study at the main campus in Atlanta, at campuses in France and China, and through distance and online learning.

As a leading technological university, Georgia Tech is an engine of economic development for Georgia, the Southeast, and the nation, conducting more than $1 billion in research annually for government, industry, and society.

Writer: Anne Wainscott-Sargent

What if there was a way for the school supplies and food to be delivered right to your dorm – not by car or foot, but by drone?

One class that is part of the Vertically Integrated Projects (VIP) Program at the Georgia Tech Research Institute (GTRI) and Georgia Tech could soon turn that idea into a reality.

The class, called Experimental Flights, is developing a drone delivery network that would allow students on Georgia Tech's campus in Atlanta to place orders for items such as school supplies and food through a mobile app, and have a drone deliver those items to a secure locker station close to their dorm. The app would have a similar look and feel to the app used for popular ridesharing services and students could use it to view wait times for the next available drone, track their package, and receive a unique code to access their purchase.

Michael Mayo, a GTRI senior research engineer who is the lead instructor for the class, said his initial goal is to roll out the drone delivery network to students at Georgia Tech and then to consider other locations later on.

"We’ve been working on this kind of network for a couple of years now and have leveraged knowledge from a lot of different disciplines at Tech – including aerospace engineering, mechanical engineering, and computer science," Mayo said. "Success for this project would be for us to develop a fully-functional drone delivery network on Georgia Tech's campus that would serve as a model for future drone delivery networks across the country and world."

VIP is an education program supported by Tech and GTRI that allows undergraduate and graduate students to earn academic credit for working with faculty on projects they don't typically encounter in a classroom setting.

Student teams work closely with faculty advisors and graduate student mentors. Classes are held once a week, though team members usually hold additional meetings outside of class. Prospective students who are interested in joining the program can apply to a team that interests them on Tech's VIP website.

Diversity of Thought

The Experimental Flights class attracts a diverse group of class years and majors.

For the spring 2022 semester, the course included 33 undergraduate students ranging from first years to fourth years with the following majors: aerospace engineering, mechanical engineering, electrical engineering, and computer science. Twenty-one of the 33 students took the class in a previous semester.

One of those students is Catherine Heaton, a fourth-year aerospace engineering major who has participated in the Experimental Flights class since the fall 2020 semester. Heaton said working with a diverse group of students has enabled her to apply the concepts she has learned from her major to solve real-world issues, while also gaining experience developing hardware systems that supports emerging technologies.

"I'm on our class' hardware team, so I help assemble all of the parts of the drone and also work a little bit with 3D software modeling," Heaton said. "There's a lot of new technologies coming out – whether it's drones, or other plane-related things – and they all have so much potential."

Another student, Tim Boyer, a third-year electrical engineering major who has also been a member of the class since fall 2020, said he most enjoys VIP's interdisciplinary focus and getting the chance to tinker with drones.

"I really enjoy working with mechanical engineering and computer science majors to make a project come together," Boyer said. "It's also great because I have always been interested in drones, so this class is a great outlet to play around with that kind of hardware."

VIP Programs are now active in over 40 universities, with more than 4,500 students participating per term around the globe. The entire Georgia Tech VIP program currently serves 84 VIP teams involving more than 200 faculty and over 1,500 students. GTRI has 13 VIP teams that involve roughly 40 faculty members.

Preparing for Launch

Mayo's class has assembled a few drone prototypes with the help of drone assembly kits and 3D printing.

The cost to create one drone is under $1,000, and each prototype can currently carry packages that weigh up to 2 pounds, according to Mayo.

"The cost of drones, batteries and other associated components continue to decrease, which makes the economics of this type of delivery system more and more favorable," Mayo said.

Drone delivery offers several benefits to traditional car-based services, including the potential for reduced greenhouse gas emissions as smaller and lighter packages are transported via drones instead of delivery trucks. This alternative delivery method could also reduce roadway congestion and lower the risk of car accidents. Drone delivery could also enable greater route flexibility, resulting in consumers receiving their packages sooner.

Beyond package delivery, drones are useful in disaster relief settings when organizations need to send goods to places with restricted access, and also in military settings to help ground troops collect key intelligence and not risking helicopter crews to deliver supplies.

The Experimental Flights class has successfully completed initial flight testing for their drones in a controlled environment that has been approved by the Georgia Tech Police Department and demonstrated the drones' ability to transport small packages. The class has also constructed a prototype package locker that can securely store multiple packages and that the drone can directly drop packages into.

The class is currently designing the mobile app for end users and a flight control center to manage drone operation. The path the drone takes through campus for each delivery will be automatically generated using an algorithm designed by the class. The algorithm has been designed to optimize the drone's flight path to ensure maximum safety by avoiding flight over people while also reducing delivery times when possible. Drones will fly themselves autonomously to their destination during normal operation.

Mayo noted a fully-operational drone would transmit real-time telemetry and live video streams to the flight control center at all times, and in the event of an emergency, a human operator would assume manual control of the drone. Packages will be secured with both an electromagnet and with the landing gear of the drone itself during transport to reduce the risk of a package becoming dislodged during flight. Rotor cowlings will be added to the drones to minimize the chance of human contact with the rotors – or a fanlike component that drones rely on for propulsion and control – during normal operation and in the event that a drone flies off its approved path.

Before implementing a drone delivery network on campus, the class would need to gain approval from campus administrators and the Federal Aviation Administration (FAA).

"Special preparation will also need to be made to get FAA approval to fly the drones beyond visual line of sight, which is a requirement for most drone operations," Mayo said.

Once the drone delivery system becomes fully operational, the only initial cost to students would be the items that they order, Mayo said. An additional delivery cost, similar to those for food delivery services such as DoorDash and Uber Eats, could be included later on.

Looking ahead, the class aims to perform flight tests where the drone would pick up a sample package and deliver the item to a locker station in one trip.

Beyond the Classroom

Mayo's class is currently seeking corporate collaborations to apply their drone delivery concept to areas such as inventory management and more widespread package delivery. His class is currently collaborating with U.S. furniture company Steelcase to study the use of drones for indoor and outdoor inventory management.

Mayo said he considers a collaboration between students and companies to be a win-win for both groups. Companies are able to build relationships with students who have in-demand skills and who could be hired as entry-level employees. Students, meanwhile, are able to receive feedback from experienced engineers and network with a company that could serve as a potential employment opportunity.

"There are so many advantages to VIP that extend well beyond the classroom," Mayo said.

Writer: Anna Akins
Photos: Christopher Moore
GTRI Communications
Georgia Tech Research Institute
Atlanta, Georgia USA

The Georgia Tech Research Institute (GTRI) is the nonprofit, applied research division of the Georgia Institute of Technology (Georgia Tech). Founded in 1934 as the Engineering Experiment Station, GTRI has grown to more than 2,800 employees supporting eight laboratories in over 20 locations around the country and performing more than $700 million of problem-solving research annually for government and industry. GTRI's renowned researchers combine science, engineering, economics, policy, and technical expertise to solve complex problems for the U.S. federal government, state, and industry.

Protein function heavily depends on their three-dimensional structure, and researchers around the world have long endeavored to answer a seemingly simple inquiry to bridge function and form: if you know the building blocks of these molecular machines, can you predict how they are assembled into their functional shape?

This question is not so easy to answer. With complex structures dependent on intricate physical interactions, researchers have turned to artificial neural network models – mathematical frameworks that convert complex patterns into numerical representations – to predict and “see” the shape of proteins in 3D.

In a new paper published in Nature Communications, researchers at Georgia Tech and Oak Ridge National Laboratory build upon one such model, AlphaFold 2, to not only predict the biologically active conformation of individual proteins, but also of functional protein pairings known as complexes.

The work could help researchers bypass lengthy experiments to study the structure and interactions of protein complexes on a large scale, said Jeffrey Skolnick, Regents’ Professor and Mary and Maisie Gibson Chair in the School of Biological Sciences and one of the corresponding authors of the study, adding that computational models such as these could mean big things for the field. 

If these new computational models are successful, Skolnick said, “it could fundamentally change the way biological molecular systems are studied.”

Primed for Protein Prediction

Created by London-based artificial intelligence lab DeepMind, AlphaFold 2 is a deep learning neural network model designed to predict the three-dimensional structure of a single protein given its amino acid sequence. Skolnick and fellow corresponding author, Mu Gao, senior research scientist in the School of Biological Sciences, shared that the Alphafold 2 program was highly successful in blind tests occurring at the 14^th iteration of the Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction, or CASP14, a bi-annual competition where researchers around the globe gather to put their computational models to the test. 

“To us, what is striking about AlphaFold 2 is that it not only makes excellent predictions on individual protein domains (the basic structural or functional modules of a protein sequence), but it also performs very well on protein sequences composed of multiple domains,” Skolnick shared. And so with the ability to predict the structure of these complicated, multi-domain proteins, the research team set out to determine if the program could go a little further. 

“The physical interactions between different [protein] domains of the same sequence are essentially the same as the interactions gluing different proteins together,” Gao explained. “It quickly became clear that relatively simple modifications to AlphaFold 2 could allow it predict the structural models of a protein complex.” To explore different strategies, Davi Nakajima An, a fourth-year undergraduate in the School of Computer Science, was recruited to join the team’s effort.

Instead of plugging in the features of just one protein sequence into AlphaFold 2 per its original design, the researchers joined the input features of multiple protein sequences together. Combined with new metrics to evaluate the strength of interactions among probed proteins, their new program AF2Complex was created.

Charting New Territory

To put AF2Complex to the test, the researchers partnered with the high-performance computing center, Partnership for an Advanced Computing Environment (PACE), at Georgia Tech, and charged the model with predicting the structures of protein complexes it had never seen before. The modified program was able to correctly predict the structure of over twice as many protein complexes as a more traditional method called docking. While AF2Complex only needs protein sequences as input, docking relies on knowing individual protein structures beforehand to predict their combined structure based on complementary shapes.

“Encouraged by these promising results, we extended this idea to an even bigger problem, which is to predict interactions among multiple arbitrarily chosen proteins, e.g., in a simple case, two arbitrary proteins,” shared Skolnick.

In addition to predicting the structure of protein complexes, AF2Complex was charged with identifying which of over 500 pairs of proteins were able to form a complex at all. Using newly designed metrics, AF2Complex outperformed conventional docking methods and AlphaFold 2 in identifying which of the arbitrary pairs were known to experimentally interact.

To test AF2Complex on the proteome scale, which encompasses an organism’s entire library of the proteins that can be expressed, the researchers turned to the Summit Oak Ridge Leadership Computing Facility, the world’s second largest supercomputing center. “Thanks to this resource, we were able to apply AF2Complex to about 7,000 pairs of proteins from the bacteria E. coli,” Gao shared. 

In that test, the team’s new model not only identified many pairs of proteins known to form complexes, but it was able to provide insights into interactions “suspected but never observed experimentally,” Gao said. 

Digging deeper into these interactions revealed a potential molecular mechanism for protein complexes that are particularly important for energy transport. These protein complexes are known to carry hemes, essential metabolites giving blood dark red color. Using AF2Complex’s predicted structural models, Jerry M. Parks, a senior research and development staff scientist at Oak Ridge National Laboratory and a collaborator in the study, was able to place hemes at their suspected reaction sites within the structure. “These computational models now provide insights into the molecular mechanisms for how this biomolecular system works,” Gao said. 

“Deep learning is changing the way one studies a biological system,” Skolnick added. “We envision methods like AF2Complex will become powerful tools for any biologist who would like to understand molecular mechanisms of a biosystem involving protein interactions.”

AF2Complex is an open-source tool available to the public and can be downloaded here.

This work was supported in part by the DOE Office of Science, Office of Biological and Environmental Research (DOE DE-SC0021303) and the Division of General Medical Sciences of the National Institute Health (NIH R35GM118039). DOI: https://doi.org/10.1038/s41467-022-29394-2