The team will consider three different approaches for creating entangled quantum memories that could facilitate the long-distance transmission of secure information. The five-year Multidisciplinary University Research Initiative (MURI) will be led by the Georgia Institute of Technology and include scientists from Columbia University, Harvard University, the Massachusetts Institute of Technology, the University of Michigan, Stanford University and the University of Wisconsin.

“We want to develop a set of novel and powerful approaches to quantum networking,” said Alex Kuzmich, a professor in Georgia Tech’s School of Physics and the MURI’s principal investigator.  “The three basic capabilities will be (1) storing quantum information for longer periods of time, on the order of seconds, (2) converting the information to light, and (3) transmitting the information over long distances. We aim to create large-scale systems that use entanglement for quantum communication and potentially also quantum computing.”

The MURI scientists will study three different physical platforms for designing the matter-light interaction used to generate the entangled photons.  These include neutral atom memories with electronically-excited Rydberg-level interactions, nitrogen-vacancy (NV) defect centers in diamonds, and charged quantum dots.

“A large body of work has been initiated in this area over the past 15 years by our team members and their research groups,” Kuzmich noted. “The physical approaches are different, but the goals are closely related, so there are significant opportunities for synergistic activities. Through this MURI, we will be able to interact more closely, communicate more quickly and provide new opportunities for our students and postdoctoral fellows.”

Overall, the MURI has four major goals:

• To implement efficient light-matter interfaces using three different approaches to entanglement;
• To realize entanglement lifetimes of more than one second in both the nitrogen-vacancy centers and atomic quantum memories;
• To implement two-qubit quantum states within memory nodes;
• To integrate different components and physical implementations into small units capable of significant quantum processing tasks.

Quantum memories generated from the interaction of neutral atoms and light now have maximum lifetimes of approximately 200 milliseconds.  But improvements beyond memory lifetime will be needed before practical systems can be created.

“We aim to be able to combine systems, so that instead of just one memory entangled with one photon, perhaps we could have four of them,” Kuzmich added.  “This may look like a straightforward thing to do, but this is not easy in the laboratory.  The improvements must be made at every level, so the difficulty is significant.”

Among the challenges ahead are maintaining separation between the different memory systems, and minimizing loss of light as signals propagate through the optical fiber systems that would be used to transmit entangled photons.  

“Light is easily lost, and there’s not much that can be done about that from a fundamental physics standpoint,” said Kuzmich.  “The rates of these protocols go down rapidly as you try to scale up the systems.”

Kuzmich and his Georgia Tech research team have been developing quantum memory based on the interaction of light with neutral atoms such as rubidium.  They have made substantial progress over the past decade, but he says it’s not clear which approach will ultimately be used to create large-scale quantum communication system.

The most immediate applications for the quantum memory are in secure communications, in which the entanglement of photons with matter would provide a new form of encryption.

“The immediate focus is on communication, including memories and distributed systems, which is important for sharing and transmitting information,” Kuzmich explained.  “It also has implications for quantum computation because similar techniques are often used.”

In addition to Kuzmich, collaborators in the MURI include:

• Luming Duan, professor of physics in the School of Physics at the University of Michigan, Ann Arbor, Michigan.
• Dirk Englund, assistant professor of electrical engineering and applied physics in the School of Engineering and Applied Science at Columbia University, New York, New York.
• Marko Lonkar, associate professor of electrical engineering in the School of Engineering and Applied Sciences at Harvard University, Cambridge, Massachusetts.
• Brian Kennedy, professor of physics in the School of Physics at the Georgia Institute of Technology, Atlanta, Georgia.
• Mikhail Lukin, professor of physics in the Department of Physics at Harvard University, Cambridge, Massachusetts.
• Mark Saffman, professor of physics in the Department of Physics at the University of Wisconsin, Madison, Wisconsin.
• Jelena Vuckovic, associate professor of electrical engineering in the Department of Electrical Engineering at Stanford University, Stanford, California.
• Vladan Vuletic, the Lester Wolfe Professor of Physics in the School of Physics at Massachusetts Institute of Technology, Cambridge, Massachusetts.
• Thad Walker, professor of physics in the Department of Physics at the University of Wisconsin, Madison, Wisconsin.

“If we are successful with this over the next five years, long-distance quantum communications may become promising for real-world implementation,” Kuzmich added.  “Integrating these advances with existing infrastructure – optical fiber that’s in the ground – will continue to be an important engineering challenge.”

This material is based upon work conducted under contract FA9550-12-1-0025.  Any opinions, findings and conclusions or recommendations expressed are those of the researchers and do not necessarily reflect the views of the Air Force Office of Scientific Research.

Research News & Publications Office
Georgia Institute of Technology
75 Fifth Street, N.W., Suite 314
Atlanta, Georgia  30308  USA

Media Relations Contacts: John Toon (404-894-6986)(jtoon@gatech.edu) or Abby Robinson (404-385-3364)(abby@innovate.gatech.edu).

Writer: John Toon

An assistant professor in the School of Electrical and Computer Engineering (ECE) at Georgia Tech, Dr. Hong was honored for his paper, "Improving Prediction Accuracy of Protein-DNA Docking with GPU Computing." He shares this award with two coauthors–Jiadong Wu, his Ph.D. student, and Jun-tao Guo, a colleague from the Department of Bioinformatics and Biomedicine at the University of North Carolina at Charlotte.

Protein-DNA docking represents one of the most challenging problems in structural bioinformatics. Knowledge of how proteins interact with DNA is critical for understanding many key biological processes and for structure-based drug design. This paper describes a high performance computing method that Dr. Hong and his team have developed to tackle the protein-DNA docking problem using a GPU cluster. This protein-DNA docking algorithm integrates Monte-Carlo simulation and a simulated annealing method and has achieved 10.4 TFLOPS of sustained performance using 128 GPU cards, which represents 4× speed up over a traditional cluster with 1000 CPU cores. Such improved computation capability accelerates the conformational space sampling for the docking algorithm and increases the chance of finding near-native protein-DNA structures.

The research, which comprises three separate projects, includes work with a new mouse model of ovarian cancer to identify early detection biomarkers; an effort to characterize proteins and protein variants secreted from ovarian tumors that could serve as serum biomarkers; and work to identify metabolic changes that could help diagnose the disease.

"This grant is a program project development grant, and the idea is to bring together a number of individuals around a common theme," Martin Matzuk, a BCM professor of pathology and immunology and one of the leaders of the project, told ProteoMonitor. "We were previously funded by OCRF along with a number of investigators to focus on the role of microRNAs in ovarian cancer. That work has gone very well, so we put together another proposal in which we decided to focus on biomarkers, whether they're protein or small molecule."

Matzuk is collaborating on the work with his BCM colleague Laising Yen as well as John McDonald, a professor, the associate dean for biology program development in the school of Biology at Georgia Tech, and a chief research scientist at Atlanta's Ovarian Cancer Institute.
McDonald, who will head up the search for metabolomic biomarkers, leads a research team that published a paper in August 2010 detailing a metabolomic ovarian cancer diagnostic that identified women with ovarian cancer with 100 percent accuracy in a 94-subject trial (PM 8/20/2010).

That test used direct-analysis-in-real-time mass spectrometry to measure thousands of metabolites in subjects' blood samples, classifying them with a functional support vector machine-based machine-learning algorithm. McDonald's team is still validating their findings, McDonald told ProteoMonitor this week, but thus far "everything is looking good," and, he said, the researchers hope to finish validating the results sometime within the year.

Under the OCRF grant, the Georgia Tech team plans to use LC-MS/MS to identify specific metabolites detected by their DART-MS work in hopes of combining them with protein biomarkers identified by Matzuk's lab to build an early detection panel for ovarian cancer.
The DART analysis "gives us thousands of features, and for most of them we don't know what they are," McDonald said. "From a diagnostic point of view we don't really care as long as it's a reliable diagnostic. But at the same time we're now running LC-MS/MS to try to whittle it down to identify … the specific metabolites involved."

"The idea is that we'll put it together [with Matzuk's markers] to see what an optimal diagnostic might consist of," he said.

Matzuk and the BCM researchers will be looking for protein biomarkers using a recently developed mouse model of high-grade serous ovarian cancer in which the cancer actually begins in the fallopian tube as opposed to the ovary itself. The model reflects an alternate view of ovarian cancer development "that is gaining a lot of support," Matzuk said.

Because ovarian cancer is difficult to detect early, often by the time patient samples are collected it's "too late to be trying to figure out what are the changes with regard to proteins or metabolic changes," he said. "The nice thing about having a mouse model is that these animals get cancers universally, and so you can open the animals up at a certain period and say, 'OK, at this time point what are the expression changes in these cancers? What are the earliest time points [they are visible]?"

"The goal of all three projects is to [identify] the various transcripts that are out there in these cancers," Matzuk said. "The idea is, once we catalog all of them, to go back in and then screen or develop antibodies to new variants of proteins or new secreted proteins and see whether or not those could be better markers."

The ultimate goal of the work, he said, "is to generate enough data so that we could actually go into the National Institutes of Health for a bigger project that we could start not only between our groups, but also with other groups and centers to look at various biomarkers."

Price will be a major consideration for any early detection test, Matzuk said, noting that he thinks even existing triage tests like Vermillion's OVA1 don't offer enough to justify their cost. Given the low prevalence of ovarian cancer in the general population, he said, any broad screening test for the disease would need to cost under $50 for it to be covered widely by insurers.

"I run a clinical chemistry laboratory in the county hospital, and for us to be doing this kind of screening of healthy women you need to have the cost low," he said.
However, Matzuk suggested, declining instrumentation prices could help bring costs down in the future – particularly in the case of mass spec-based tests, where multiplexing could significantly lower the price of multi-analyte assays.

"Maybe everyone will have [mass spec] analysis of their serum at some point," he said. "I think right now the instrumentation is too expensive and the testing is too expensive to go ahead and say this is for general [screening] tests, but if it turns out that these tests are extremely valuable, people are going to find a way to make them cheaper."

From 2008 to 2010, a team of bioinformatics graduatestudents, led by School of Biology Associate Professor King Jordan, worked inclose collaboration with the Centers for Disease Control and Prevention (CDC)to create an integrated suite of computational tools for the analysis ofmicrobial genome sequences.  At thattime, CDC scientists were in need of a fast and accurate system that couldautomate the analysis of sequenced genomes from disease-causing bacteria. Theyturned to the Jordan lab at Georgia Tech to help develop such a tool. TheGeorgia Tech scientists created an open source software package, the ComputationalGenomics Pipeline (CG-pipeline), to help meet CDC’s need. The softwareplatform is now used worldwide in public health research and response efforts.

“Determining the order of DNA bases for an entire genome hasbecome relatively cheap and easy in recent years because of technologicaladvancements,” said Jordan. “The hard part is figuring out what the genomesequence information means. Our software takes that next step. It analyzes the sequences,finds the genes and provides clues as to which genes are involved in makingpeople sick. Manually, this process used to take weeks, months or a year. Nowit takes us about 24 hours.”

The CG-pipeline software has been used to analyze lastsummer’s outbreak of  severe Escherichia coli(E. coli) infections that started in Germany and eventually led to illnesses in16 European countries, Canada and the United States. It was one of the largest E.coli outbreaks in history, causing 50 deaths and 4,075 confirmed worldwidecases. The bacterium was traced to sprouts. Andrey Kislyuk, a graduate of the BioinformaticsPh.D. program who helped Jordan create the software, used the CG-pipeline whileworking at Pacific Biosciences to understand why the strain of the bacteriathat caused the outbreak was so virulent.

“The software was used to determine that genetic materialfrom two previously distinct strains of E. coli  was combined in a new, hyper-virulent strain,” said Kislyuk. “Theresulting hybrid strain seems to be more lethal than either of the parentstrains.”

Another Bioinformatics Ph.D. graduate who helped design andimplement the pipeline, Lee Katz, analyzed the bacteria that caused last year’soutbreak of listeriosis in the United States while working at the CDC.  That outbreak was traced back to cantaloupesfrom a single farm in Colorado that were tainted with Listeria. Over the spanof several months, there were 146 confirmed cases of listeriosis and 30 deaths,making it the deadliest outbreak of foodborne illness in the U.S. in 25 years.Using the CG-pipeline, Katz was able to identify an important epidemiological genomicmarker, which will help track invasive strains of Listeria.

The CG-pipeline software platform can be used to analyze anymicrobial genome sequence. It has already been applied to bacteria that cause avariety of infectious diseases, including cholera, salmonella and bacterialmeningitis.

Katz continues to work closely with the Jordan lab toimprove the software. This collaboration is important in CDC’s efforts to minegenome sequence information in the service of public health using softwaredeveloped at Georgia Tech.

“During Dr. Yuan’s time on the ISyE faculty, he has made valuable contributionsin research, teaching, and service. Dr. Yuan’s exceptional teaching ability isevident in the excellent teaching evaluations and student praise hereceives.  We are fortunate to have him as a colleague and now as theISyE’s newest Coca-Cola Junior Professor,” said Jane C. Ammons, H. Milton andCarolyn J. Stewart School Chair.

The Coca-Cola Junior Professorship is supported by a gift from Coca-Cola, inorder to support research and development in ISyE. Endowed professorships, suchas this one, are awarded to outstanding faculty, ensuring them the resourcesthey need to remain at the forefront of their fields and to lead teaching andresearch efforts in their key areas.

In addition to this recent honor, Yuan was the recipient of the NationalScience Foundation Career Award in 2009 for his exemplary work in sparsemodeling and estimation with high-dimensional data.  He was also named asa Distinguished Cancer Scholar from the Georgia Cancer Coalition in 2007, andwas the recipient of the John van Ryzin Award in 2004.

Yuan received his PhD in statistics from the University of Wisconsin atMadison. He also holds a master’s in computer science from the University ofWisconsin, and a bachelor’s in electrical engineering and information sciencefrom the University of Science & Technology of China. Yuan's currentresearch interests include statistical learning, bioinformatics, and methods ofregularization.

The following awards were presented to IAC faculty:

Big Data Award
Alan L. Porter, Professor Emeritus, Schools of Public Policy and Industrial & Systems Engineering

The Big Data Award recognizes a Georgia Tech researcher or research group that has successfully established a strong research relationship with industry and commercialized technology for the management and application of large complex datasets to solve problems in society and industry and, in so doing, developed new tools and methods for capture, storage, analysis, searching and visualization of information.

People and Technology Award
Michael L. Best, Associate Professor, The Sam Nunn School of International Affairs and School of Interactive Computing; Editor-in- Chief, Information Technologies & International Development

The People and Technology Award honors the efforts of Georgia Tech researchers whose pursuit of research in the interaction between humans and technology demonstrates a transformative societal impact in improving the human condition.

Public Service, Leadership, and Policy Award
Dan Breznitz, Associate Professor, The Sam Nunn School of International Affairs

The Public Service, Leadership, and Policy Award recognized a Georgia Tech researcher or research group that has made a significant contribution to the study of innovation and policies that promote and sustain research for public benefit.

Innovation in Literature and Communication Award
Thomas N. Lux, Professor, School of Literature, Communication, and Culture; Founder and Director of Poetry@Tech

The Innovation in Literature and Communication Award is given in recognition for innovation in communication and literature in science, engineering, and technology.

The Gala Celebration Honoring Innovators and Inventors & GTRC for 75 years of Service to Georgia Tech Faculty took place December 12, 2011 at the Georgia Tech Hotel and Conference Center Ballroom.

The NMSC focused on establishing, for the first time, a national agenda for maintaining the growth of modeling and simulation technology and its incorporation into all areas of the national economy, welfare, and security.

Richard Fujimoto, chair of the School of Computational Science and Engineering and interim director of the Institute for Data and High Performance Computing at the Georgia Institute of Technology, is a member of the interim Board of Directors and interim chair of the Education and Professional Development Standing Committee, one of five standing committees of the National Modeling and Simulation Coalition. He chaired the inaugural meeting for this committee at the Feb. 6 event.

“A critical issue that the committee is starting to address is how to increase the number of people trained and educated in modeling and simulation in order to meet the high workforce demand,” Fujimoto said. “We are developing a national agenda that will focus in part on this issue.”

The committee’s work will span the entire education pipeline, including K-12, higher education, technical education and continuing education.

Within the School of Computational Science and Engineering at Georgia Tech, the approach is to develop education programs in computing for students – who are studying engineering, the sciences and other areas - that build up their computing capability, said Fujimoto.

“We really must be focused on developing workforce needs,” the CSE chair said.

At the event in Washington, D.C., Aneesh Chopra, chief technology officer for the United States, gave the keynote address after a ceremonial ribbon cutting for the NMSC.

The event featured several hundred key leaders from industry, government, and academia. Also attending from Georgia Tech was Margaret Loper, a chief scientist in the Georgia Tech Research Institute.

Participants set the stage for the initial organizational and committee sessions and to define a detailed action plan in four key areas: Education and Professional Development, Technology Research and Development, Industrial Development, and Business Practice. The outcome of the inaugural meeting included a national action plan to broaden the use of modeling and simulation across these four key areas. This action plan will provide a road map for the modeling and simulation community and in expanding the depth and breadth of the technology and industry.

Finally, in 2011, Shechtman earned the ultimate vindication, with a Nobel of his own.

In a standing-room-only lecture on Feb. 23 at Georgia Tech, Shechtman told the story of his 1982 discovery of quasi-periodic crystals, which went against everything that was known about the structure of crystals and resulted—30 years and many scientific battles later—in his being awarded the 2011 Nobel Prize in Chemistry.

CLICK HERE FOR VIDEO OF THE LECTURE

Shechtman, who is Philip Tobias Professor of Materials Science at the Technion – Israel Institute of Technology (click here for full bio and abstract), laid the groundwork for his audience by explaining that modern crystallography began in 1912 with the seminal work of German physicist Max von Laue. Von Laue established the three basic principles of crystalline structure: order, periodicity and rotational symmetry. Shechtman then carefully defined each term, concluding with the universally accepted—until Shechtman’s discovery—definition of crystals as “solids composed of atoms arranged in a pattern that is periodic and in three dimensions,” and that the rotational symmetry of crystals could be one-, two-, three-, four- or six-fold—never five-fold and never more than six.

But in 1982 as a young faculty member at the Technion, Shechtman one day discovered an apparent tenfold crystalline rotational symmetry in a crystal composed of aluminum and manganese, and further found that the crystal’s structure was not periodic but quasi-periodic. To drive home his own surprise at the time, Shechtman displayed an image of the actual page from his 30-year-old notebook, with “10fold!!!” written clearly by a set of notations.

“At the end of the day, I knew this was something new and exciting,” he said, further adding that years earlier in graduate school he’d been given a test in which he had to prove that such rotational symmetry in crystals was impossible. “And I did it. I passed the test. I would not be here [talking to you] if I hadn’t.”

So began a decade-long odyssey of Shechtman persisting to convince more and more scientists of his findings. First it was his group research leader at the U.S. National Bureau of Standards (now called the National Institute of Standards and Technology), who called Shechtman a “disgrace” and kicked him out of the group. Then it was members of the International Union of Crystallography (IUCr), who rejected the study due to Shechtman’s having used an electron microscope; the IUCr insisted that legitimate studies of crystalline diffraction used X-rays, not electron beams.

Finally it came down to one man: Linus Pauling, winner of Nobel Prizes in two different fields and one of the greatest scientists of the 20th century. Pauling rejected the idea of quasi-periodic crystals right up until his death in 1994, and envious colleagues used Pauling’s objections to argue against Shechtman’s academic promotions, even as crystallographers around the world came to accept his work.

In the end, of course, the discovery was accepted, and Shechtman proudly pointed to a new definition of crystals, adopted in 1991, that acknowledges “aperiodic crystals … in which three-dimensional lattice periodicity can be considered to be absent.”

“It is a soft, humble definition,” he said. “And a humble scientist is a good scientist.”

Shechtman’s visit was sponsored by the Georgia Tech colleges of Computing, Science and Engineering, as well as the Georgia Tech Executive Vice President for Research, the Georgia Tech Research Institute, the Georgia Tech Institute for Leadership and Entrepreneurship, the American-Israel Chamber of Commerce-Southeast Region, the American-Israel Educational Institute, and Given Imaging.

To address that need, systems engineers at the Georgia Institute of Technology are using computer models to help resource-poor nations improve supply chain decisions related to the distribution of breast milk and non-pharmaceutical interventions for malaria. They are also forecasting what health care services would be available in the event of natural disasters in Caribbean nations.

“We are using mathematical models implemented in user-friendly tools like Microsoft Excel to improve the allocation of limited resources across a network, especially in resource-poor settings,” said Julie Swann, an associate professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech.

Swann reported on three global health case studies designed to improve the allocation of limited health care resources on Feb. 19, 2012 at the annual meeting of the American Association for the Advancement of Science (AAAS) in Vancouver, Canada.

For the first project, Swann and a group of graduate students created models to strategically determine how a nongovernmental organization (NGO) in South Africa should expand its breast milk donation and distribution network to the whole country. In the network, healthy mothers donate breast milk, which is stored in a local repository, transferred to a milk bank to be processed and then distributed to neonatal units where mothers cannot provide it themselves because of disease status or physical inability.

“We wanted to determine how we could provide breast milk to the most people while also being geographically equitable in terms of access,” explained Swann, who holds the Harold R. and Mary Ann Nash chair at Georgia Tech. “We looked at the cost of equity and how that changed the distribution design.”

To determine where the organization should expand its network and the best way to do so, the team used operations research to examine the existing and proposed locations in the network as well as what type of transportation would work best to cover the increased geographic area. The model recognized that breast milk supply increases with higher income and education levels and low HIV prevalence, while breast milk demand increases with lower income and education levels and high HIV prevalence.

The researchers recently recommended locations for expansion to the NGO and advised the organization to pay a courier service to carry the milk to the neonatal units, in order to balance cost and reliability and improve efficiency. Volunteers, who are inherently less reliable, were driving the milk from one location to another.

In another project, done in collaboration with the World Health Organization, Swann and a team of undergraduate and graduate students used models to optimize the distribution of non-pharmaceutical interventions for malaria, such as nets or sprays, with pilot data from a country in Africa called Swaziland.

Their models provided a time-based deployment plan for the country, including details on what geographic zones to target for spraying, when to deploy in each zone, how many people can be protected in each zone, what resources should be located at the distribution centers, and the opening and closing dates of the distribution centers.

The researchers showed that using a systems approach to examine allocation decisions could increase the number of people covered with the same amount of funding by more than 25 percent. The team worked with Pinar Keskinocak, a professor in the Stewart School of Industrial and Systems Engineering at Georgia Tech, to develop a teaching game based on the work. The game has been used worldwide in classes of humanitarian students.

For the third project, Swann and a team of graduate students are using technology to estimate the performance of disaster preparedness plans in advance of an event. The project is part of the Caribbean Hazard Assessment Mitigation and Preparedness (CHAMP) initiative, which is supported by a Georgia Tech alumnus and led by Reginald DesRoches, a professor in the School of Civil and Environmental Engineering at Georgia Tech.

In Puerto Rico, Swann’s team evaluated the existing hospital networks and other health care provider locations described in the island’s emergency preparedness plans.

“To forecast the country’s ability to provide health services following an earthquake, we took population data and overlaid it with projections of earthquake locations and severity to estimate the capacities and amount of congestion that would result at health care facilities,” said Swann.

The researchers recently presented the initial results of their study to the Puerto Rico Department of Health and made recommendations for health care resources and hospital capacities based on predicted bottlenecks in the system. They are currently examining Belize’s hurricane evacuation plans. Keskinocak and Stewart School of Industrial and Systems Engineering associate professor Ozlem Ergun and visiting assistant professor Pelin Pekgun-Cakmak are also contributing to the CHAMP initiative.

“We have found that technology innovations like mathematical models can help to solve problems in global and public health, such as the allocation of limited health care resources,” noted Swann. 

Research News & Publications Office
Georgia Institute of Technology
75 Fifth Street, N.W., Suite 314
Atlanta, Georgia 30308 USA

Media Relations Contacts: Abby Robinson (abby@innovate.gatech.edu; 404-385-3364) or John Toon (jtoon@gatech.edu; 404-894-6986)

Writer: Abby Robinson

This plethora of information is allowing scientists to share their findings in unique ways. Zhigang Peng, associate professor in Georgia Tech’s School of Earth and Atmospheric Sciences, has converted the earthquake’s seismic waves into audio files. The results allow experts and general audiences to “hear” what the quake sounded like as it moved through the earth and around the globe.

“We’re able to bring earthquake data to life by combining seismic auditory and visual information,” said Peng, whose research appears in the March/April edition of Seismological Research Letters. “People are able to hear pitch and amplitude changes while watching seismic frequency changes. Audiences can relate the earthquake signals to familiar sounds such as thunder, popcorn popping and fireworks.”

The different sounds can help explain various aspects of the earthquake sequence, including the mainshock and nearby aftershocks. For example, this measurement was taken near the coastline of Japan between Fukushima (the nuclear reactor site) and Tokyo. The initial blast of sound is the 9.0 mainshock. As the earth’s plates slipped dozens of meters into new positions, aftershocks occured. They are indicated by “pop” noises immediately following the mainshock sound. These plate adjustments will likely continue for years.

As the waves from the earthquake moved through the earth, they also triggered new earthquakes thousands of miles away. In this example, taken from measurements in California, the quake created subtle movements deep in the San Andreas Fault. The initial noise, which sounds like distant thunder, corresponds with the Japanese mainshock. Afterwards, a continuous high-pitch sound, similar to rainfall that turns on and off, represents induced tremor activity at the fault. This animation not only help scientists explain the concept of distant triggering to general audiences, but also provides a useful tool for researchers to better identify and understand such seismic signals in other regions.

The human ear is able to hear sounds for frequencies between 20 Hz and 20 kHz, a range on the high end for earthquake signals recorded by seismometers. Peng, graduate student Chastity Aiken and other collaborators in the U.S. and Japan simply played the data faster than true speed to increase the frequency to audible levels. The process also allows audiences to hear data recorded over minutes or hours in a matter of seconds.

The research is published in the March/April edition of Seismological Research Letters.

For more on the anniversary of the Japan disaster, visit www.gatech.edu/experts/japan-anniversary.

This project was supported in part by the National Science Foundation (NSF) (CAREER Award No. EAR-0956051). The content is solely the responsibility of the principal investigators and does not necessarily represent the official views of the NSF.

Georgia Tech Associate Professor Hermann Fritz and his researchteam are studying the impact of the tsunami on the Sanriku coast. Usingeyewitness video and terrestrial laser scanners from atop the highest buildingsthat survived the tsunami, Fritz has mapped the tsunami’s height and flood zoneto learn more about the flow of the devastating currents.

Fritz’s measurements and observations could produce floodingforecasts that influence future evacuation plans and building designs, preventingloss of life and property damage in Japan and in other areas of the worldsusceptible to tsunamis.

“The ultimate goal is to save lives,” Fritz said. “In orderto do so, we have to have a better understanding of what worked and didn’twork. This is the first time we’ve been able to look at the structuralinfrastructure designed to protect coastal towns from tsunamis and examine whyit didn’t work. There’s a lot to learn in terms of surviving tsunamis andprotecting, evacuating and ultimately saving lives.”

Fritz led a reconnaissance team surveying the impact of thetsunami on a fishing town in Kesennuma Bay, where 1,500 people perished. Thebay has been hit by historic tsunamis in 1896, 1933, 1960 and 2010—making itthe most vulnerable in Japan to both near- and far-field tsunamis. The coastalstructures and other mitigation measures on the coast were designed based onconservative, historic high-water marks, rather than probable maximum tsunamis.

From two atop vertical evacuation buildings where eyewitnessesgathered during the tsunami, Fritz and his team used lasers to scan the portand bay entrance, creating a three-dimensional, topographic model of the floodzone.

Using this data, they reconstructed eyewitness videos todetermine the varying heights and flow velocities of the tsunami. Theydetermined that the tsunami reached a maximum height of 9 meters, followed byoutflow currents of 11 meters per second less than 10 minutes later – a speedwhich Fritz says is impossible to survive or navigate by vessels.

“What we can learn from the hydrograph is confirmation thatthe water goes out first, drawing down to more than negative 3 meters on thelandward side of the trench, which can make vessels hit ground inside harbors,”Fritz said. “During the subsequent arrival of the main tsunami wave, the waterrushing back in changed the water level by 40 feet, engulfing the entire cityin 12 minutes.”

Understanding tsunami impacts will help prepare for futuredisasters—whether its designing buildings high enough to serve as verticalevacuation points or sea walls and breakwaters strong enough to control theflow of water.

Along with such mitigation measures, Fritz says educatingpeople about tsunamis is key. 

“Japan was probably the best prepared for a tsunami,” Fritzsaid. “Indonesia, on the other hand, had no knowledge of tsunamis and it caughtpeople by surprise in 2004. The outcomes of the tsunamis were verydifferent—200,000 killed versus 20,000 killed. That shows educational awarenessand preparedness and civil defense mechanisms can work to reduce the deathtoll. People need to be tsunami-aware.”

Fritz worked with researchers from the University ofSouthern California and Japanese researchers from the University of Tokyo, theTokyo University of Marine Science and Technology, and the Port and AirportResearch Institute, in coordination with the UNESCO-organized InternationalTsunami Survey Team and the Tohoku University in Sendai.

This project was supported in part by the National ScienceFoundation (NSF) (Award No. 1135768). The content is solely the responsibilityof the principal investigators and does not necessarily represent the officialviews of the NSF.

For more on the anniversary of the Japan disaster, visit www.gatech.edu/experts/japan-anniversary.

“By being part of the SDAV team, the Georgia Tech researchers and the software artifacts we have been producing can more widely affect research around the nation,” said Schwan, director of CERCS. “More importantly, our work can enable scientists to carry out their science mission more effectively."

The Georgia Tech team is also part of one of the CoDesign projects being undertaken by DOE researchers that will work to improve the combustion processes used in internal combustion engines.

President Obama announced the $200 million Big Data Research and Development Initiative on Thursday, March 29. Visit here to learn more about the announcement.

Kernel summations are a ubiquitous key computational bottleneck in many data analysis methods. The paper proposes a hybrid MPI/OpenMP kernel summation framework for scaling many popular data analysis methods. Advantages to the approach include utilizing the platform-independent C++ code base that utilizes standard protocols such as MPI and OpenMP; using the template code structure that uses any multidimensional binary trees and any approximation schemes that may be suitable for high-dimensional problems; and having extendibility to a large class of problems that require fast evaluations of kernel sums.

“Researchers have previously parallelized kernel summations in the context of simulations,” says Dongryeol Lee, a Ph.D. candidate in Computer Science. “But this paper is the first serious effort in parallelizing kernel summations in the context of data mining with potentially high-profile scientific applications.”

In data mining, kernel summations appear in popular so-called kernel methods which can model complex, nonlinear structures in data. The richer expressiveness of the methods comes with the drawback of requiring many data points and hence more computational power for crunching collected data, according to Lee. The collected data in some cases must be stored on multiple machines.

From the data mining community, Lee says this work is the first to utilize algorithmic techniques in both high performance computing, computer science, computational physics, computational geometry, and approximation theory in a general framework.

Kernel summations drive algorithms in application areas such as finance, astronomy, and medical science. 

Lee notes some examples: “Fraudulent financial transactions can be detected more quickly using fast kernel summations. Astronomy uses the algorithms to predict redshift of many galaxies and stars, which can shed light onto the ultimate fate of the universe. Medicine uses fast kernel summation algorithms in automated early detection of cancer that can save human lives."

Research in computational science and engineering at Georgia Tech spans many areas ranging from the development of new computational methods that may be applied to one or more fields in science and engineering to novel computational approaches specific to a particular domain such as biology or aerospace engineering.

Because Georgia Tech views computation as the driver of future advances in science and engineering, the School of Computational Science and Engineering was created to be a truly interdisciplinary unit that crosses the conventional academic boundaries found between research disciplines. Faculty from all walks of computing, sciences, and engineering collaborate within six core areas: High-Performance Computing; Data Analytics, Machine Learning and Visualization; Modeling and Simulation; Computational Mathematics; Computational Science; and Computational Engineering.

The HPC500 is comprised of a representative cross-section of academic, government, and commercial organizations across all budgets, applications, and geographic areas, including users in both High Performance Technical Computing (HPTC) and High Performance Business Computing (HPBC). The charter members are listed at the HPC500 Website (http://www.hpc500.com/member-directory/).

Of the first fifty members:

• 20 are commerical organizations (13 with HPTC application, 7 HPBC), 19 are academic or non-for-profit, and 11 are government organizations.
• 25 are based in the U.S. or Canada; 14 are based in Europe, Middle East, or Africa (EMEA); nine are based in Asia Pacific (including Japan, Australia, and New Zealand); and two are based in Latin America (including Mexico).
• Nine have supercomputing budgets of over $5.0 million per year; 15 have high-end HPC budgets of $1.0 million to $4.9 million per year; 15 have mid-range HPC budgets of $100,000 to $999,999 per year, and 11 have entry-level HPC budgets under $100,000 per year.

For more information about High Performance Computing at Georgia Tech, please contact David A. Bader, professor in the School of Computational Science and Engineering and executive director of High Performance Computing, at 404-894-5756.

In computational terms, the air traffic controller is the “betweenness centrality,” the most connected person in the system. In this example, finding the key influencer is easy because each departure process is nearly the same.

Determining the most influential person on a social media network (or, in computer terms, a graph) is more complex. Thousands of users are interacting about a single subject at the same time. New people (known computationally as edges) are constantly joining the streaming conversation. 

Georgia Tech has developed a new algorithm that quickly determines betweenness centrality for streaming graphs. The algorithm can identify influencers as information changes within a network. The first-of-its-kind streaming tool was presented this week by Computational Science and Engineering Ph.D. candidate Oded Green at the Social Computing Conference in Amsterdam.

“Unlike existing algorithms, our system doesn’t restart the computational process from scratch each time a new edge is inserted into a graph,” said College of Computing Professor David Bader, the project’s leader. “Rather than starting over, our algorithm stores the graph’s prior centrality data and only does the bare minimal computations affected by the inserted edges.”

In some cases, betweenness centrality can be computed more than 100 times faster using the Georgia Tech software. The open source software will soon be available to businesses.

Bader, the Institute’s executive director for high performance computing, says the technology has wide-ranging applications. For instance, advertisers could use the software to identify which celebrities are most influential on Twitter or Facebook, or both, during product launches.

“Despite a fragmented social media landscape, data analysts would be able to use the algorithm to look at each social media network and mark inferences about a single influencer across these different platforms,” said Bader.

As another example, the algorithm could be used for traffic patterns during a wreck or traffic jam. Transportation officials could quickly determine the best new routes based on gradual side-street congestion.

The accepted paper was co-authored by Electrical and Computer Engineering Ph.D. candidate Rob McColl.

This project is supported by the National Science Foundation (NSF) (Award Number CNS-0708307). The content is solely the responsibility of the principal investigators and does not necessarily represent the official views of the NSF.

Awarded under DARPA’s Power Efficiency Revolution for Embedded Computing Technologies (PERFECT) program, the negotiated cooperative agreement contract (with options out to five years) is one piece of a national effort to increase the computational power efficiency of "embedded systems" by 75-fold over the best current computing performance in areas extending beyond traditional scientific computing. Professor David Bader, executive director of high-performance computing in the School of Computational Science & Engineering, is principal investigator on the Georgia Tech cooperative agreement, along with research scientist and co-PI Jason Riedy.

“Power efficiency is one of the greatest challenges confronting the designer of any computing system, much less one that’s capable of this kind of speed,” Bader said. “We could build this system today, but it would require megawatts of electricity—enough to power a medium-sized city. Our goal is to deliver the same graph analytic capabilities on platforms that require only watts or kilowatts.”

Such a system would have benefits in energy conservation, of course, but it could also save lives. The tactical advantages of supercomputing in military situations—quickly and comprehensively mapping individual or group social-media activity, for example—are becoming more critical every day, and the capacity simply doesn’t exist to deliver massive amounts of data from the field to a central computing system. Georgia Tech’s objective is to bring supercomputer graph-analysis capabilities where they're needed, from vehicles to field hospitals and beyond. The project bears the acronym GRATEFUL: “Graph Analysis Tackling power-Efficiency, Uncertainty and Locality.”

In addition to power efficiency, the second priority is to maximize computational resiliency, meaning the product algorithms will be able to withstand errors at the application and even hardware level that could result from input error or environmental factors (such as weather and hardware damage).

Bader and Riedy’s task is to develop the algorithmic framework upon which these new embedded systems will operate, and they will consciously remain “architecture-agnostic” so that the end product can be applied as widely as possible. Finally, like all programs funded under DARPA PERFECT, research and testing will be done in simulation rather than on actual embedded systems. GRATEFUL will be broken up into three stages: research & startup (18 months), risk mitigation (18 months) and prototyping (two years).

“Our goal is to make sure we have graph-analysis algorithms that can manage issues across architectures,” Riedy said. “And we’ll be looking at all the issues that concern hardware designers. Today's platforms maximize the number of operations running at once, while these new platforms consider the most power-efficient levels of that concurrency. These are not new concerns, but our job is to find new ways to deal with them.”

###

Contacts

Michael Terrazas

Assistant Director of Communications

College of Computing at Georgia Tech

mterraza@cc.gatech.edu

404-245-0707

“Keeneland provides an important capability for the NSF computational science community,” says Jeffrey Vetter, Principal Investigator and Project Director, with a joint appointment to Georgia Tech's College of Computing and Oak Ridge National Laboratory. “Many users are running production science applications on GPUs with performance that would not be possible on other systems.”

Scientists will be able to use the resource to create breakthroughs in many fields of science. For the past 20 months, the Keeneland Initial Delivery System (KIDS) has been used for research in both computer science and computational science, and has included applications in astronomical sciences, atmospheric sciences, behavioral and neural sciences, biological and critical systems, materials research and mechanical and structural systems, along with many other application areas. Much of the research will continue on KFS.

Keeneland’s early users note how the system’s capabilities have significantly advanced their research application areas.

“The Infiniband communication is now fast enough so that I can run my program on more GPUs to achieve better performance,” says Jens Glaser, a post-doctoral associate in chemical engineering and materials science at the University of Minnesota. Glaser believes his research results demonstrate that the KFS' hardware is a significant step forward in supercomputing.

Astrophysics researcher Jamie Lombardi, an associate professor in the Department of Physics at Allegheny College, says Keeneland is easily the fastest system he has used. Lombardi uses his hydrodynamics code Starsmasher to simulate the collision and merger of two stars. The dynamics of the gas are parallelized on the CPU cores, while the gravity calculations are parallelized on the GPUs. 

“Running on one node of KFS is nearly a factor of three faster than running on one node of my local cluster,” says Lombardi. “The availability of such a large number of nodes on KFS makes it possible for me to run higher resolution simulations than I have ever run before.”

The Keeneland Full Scale System is a 615 TFLOPS HP Proliant SL250-based supercomputer with 264 nodes, where each node contains two Intel Sandy Bridge processors, three NVIDIA M2090 GPU accelerators, 32 GB of host memory, and a Mellanox InfiniBand FDR interconnection network. KFS has delivered sustained performance of over a quarter of a PetaFLOP (one quadrillion calculations per second) in initial testing. The system is space efficient in that it occupies about 400 square feet, including the space for in-row cooling and service areas.

During the KFS installation and acceptance testing, the initial delivery system, KIDS, was used to start production capacity for XSEDE users seeking to run their applications on the system and who had received allocations for Keeneland through a peer review process. KIDS was upgraded with newer GPUs and used for software and application development and for pre-production testing of codes that utilize the GPU accelerators in the Keeneland systems. Even before KFS began production, allocation requests for time greater than the total available for its lifecycle had been received from XSEDE application users.

“Our Keeneland Initial Delivery system has hosted over 130 projects and 200 users over the past two years,” says Vetter. “Requests for access to Keeneland have far outstripped the planned resource delivery, sometimes by as much as twice the availability.”

The Keeneland Project is a five-year Track 2D cooperative agreement, which was awarded by NSF under Contract OCI-0910735 in 2009 for the deployment of an innovative high performance computing system to the open science community. The Georgia Institute of Technology, University of Tennessee-Knoxville, the National Institute for Computational Sciences, and Oak Ridge National Laboratory manage the facility, perform education and outreach activities for advanced architectures, develop and deploy software tools for this class of architecture to ensure productivity, and team with early adopters to map their applications to Keeneland architectures.

To learn more about Keeneland or XSEDE, visit http://keeneland.gatech.edu or https://www.xsede.org/, respectively.

###

Contacts

Joshua Preston

Communications Officer

College of Computing at Georgia Tech

jpreston@cc.gatech.edu

678-231-0787

Launched in 2011 to international acclaim, the Holiday Gift Guide is fast becoming a yuletide staple around the College’s halls, as faculty and students spend the year busily hammering together parallel algorithms and 50-amp servos in anticipation of the Big Day (i.e., the day of this press release). Now, with the days growing shorter and the sound of jingle bells in the air, all the gifts are wrapped, peer-reviewed and waiting for that next lucky computing aficionado.

“At Georgia Tech, we truly believe that computing is making the world a better place, so what better time of year to share some of our more exciting and beneficial research projects?” said Dean Zvi Galil. “When you take beloved holiday traditions and you add a layer of computation, they become so much more. In this case, they become a bit funnier. Or at least we hope so.”

Projects include:

• Nerdherder: A motion-controlled, augmented-reality puzzle game for mobile devices. The action literally leaps from the game board to your phone or tablet in this game out of Professor Blair MacIntyre’s lab.
• C4G BLIS: Open-source software system to track medical records and samples, and one of the more widely implemented projects to emerge from the College’s Computing for Good (C4G) initiative, headed up by Professor Santosh Vempala.
• “Dream. Encode.”: Inspirational short film that tells the story of a young girl discovering how and where to pursue her computational dreams. Directed by 2012 graduate Connie Chen.
• Keeneland Supercomputing System: Now the National Science Foundation’s fastest dedicated supercomputer for scientific research. Built by Professor Jeffrey Vetter.
• MOOCs: All the rage this year, MOOCs are massively open online courses,and they are in the midst of transforming education delivery, with Georgia Tech helping to lead the way.
• Bobble: Chrome plugin that allows users to escape the “filter bubble” created by personalized search results. Created by Ph.D. student Xinyu Xing.
• BrailleTouch: Software that allows you to go eyes-free when typing on a smartphone. Just another revolutionary advance in HCI from the folks at the GVU Center.
• Flashpoint: A crash course in the scientific way to get startups off the ground, running—and funded. Conceived and run by Professor Merrick Furst.
• SonLab: Proof that the universe is filled with music, this Georgia Tech lab takes natural data points and turns them into song. Created by Professor Bruce Walker.
• Betweenness Centrality Algorithm: The fastest algorithm for determining the most popular point on a social graph, created by Professor David Bader.
• Computing Summer Camps: Fun summer camps that will get kids of all ages engaged and interested in learning about computing.
• MacGyver Bot: Humanoid robot, created by Professor Mike Stilman, that soon will be able to create tools from objects in its environment.

Visit the 2012 Holiday Gift Guide on the College of Computing website!

DISCLAIMER: The 2012 Holiday Gift Guide is a lighthearted way to call attention to the College’s research. Though some of the items described in the Gift Guide are indeed available for purchase or free download, it is not intended as a practical reference for consumers.

###

About the Georgia Tech College of Computing

The Georgia Tech College of Computing is a national leader in the creation of real-world computing breakthroughs that drive social and scientific progress. With its graduate program ranked 10th nationally by U.S. News and World Report, the College’s unconventional approach to education is defining the new face of computing by expanding the horizons of traditional computer science students through interdisciplinary collaboration and a focus on human-centered solutions. For more information about the Georgia Tech College of Computing, its academic divisions and research centers, please visit http://www.cc.gatech.edu.

Contacts

Brendan Streich

Director of Communications

College of Computing at Georgia Tech

bstreich@cc.gatech.edu

404-894-7253

Below is a synopsis of Georgia Tech researchers participating and their research topics. Details can be found at the technical program schedule.

Georgia Tech @ SIAM CSE13

Panel Discussion: Big Data Meets Big Models

Panelist: David A. Bader

Minisymposium: Frontiers in Large-Scale Graph Analysis

Organizers: Jason Riedy, Henning Meyerhenke, David A. Bader

Talks:

MS25

Classifying Soft Error Vulnerabilities in Extreme-Scale Scientific Applications Using Bifit

Authors: Jeffrey S. Vetter, Dong Li (Oak Ridge), Weikuan Yu (Auburn U)

MS141

Applications and Challenges in Large-scale Graph Analysis

Authors: David A. Bader, Jason Riedy, Henning Meyerhenke

MS152

Large-scale Biomolecular Electrostatics with Massively Parallel FMM

Author: Aparna Chandramowlishwaran

MS179

Analyzing Graph Structure in Streaming Data with STINGER

Authors: Jason Riedy, David A. Bader, Robert C. McColl, David Ediger

MS195

Tensor Hypercontraction Theory: A Physically-Motivated Rank Reduction Method for Electronic Structure Theory

Author: Robert M. Parrish

MS199

On the Consistency of Calibration Parameter Estimation in Deterministic Computer Experiments

Author: Jeff Wu

MS225

PASQUAL: Parallel Techniques for Next Generation Genome Sequence Assembly

Authors: Xing Liu, Pushkar Pande, Henning Meyerhenke, David A. Bader

MS231

A ’Roofline’ Model of Energy and What it Implies for Algorithm Design

Author: Jee Whan Choi, Rich Vuduc

One proposed solution is to divide the computing among multiple small quantum computers that would work together much as today’s multi-core supercomputers team up to tackle big digital operations. The individual computers in such a system could communicate quantum information using Bose-Einstein condensates (BECs) – clouds of ultra-cold atoms that all exist in exactly the same quantum state. The approach could address the decoherence problem by reducing the number of bits necessary for a single computer.

Now, a team of physicists at the Georgia Institute of Technology has examined how this Bose-Einstein communication might work. The researchers determined the amount of time needed for quantum information to propagate across their BEC, essentially establishing the top speed at which such quantum computers could communicate.

“What we did in this study was look at how this kind of quantum information would propagate,” said Chandra Raman, an associate professor in Georgia Tech’s School of Physics. “We are interested in the dynamics of this quantum information flow not just for quantum information systems, but also more generally for fundamental problems in physics.”

The research is scheduled to be published in the April 19 online version of the journal Physical Review Letters. The research was funded by the U.S. Department of Energy (DOE) and the National Science Foundation (NSF). The work involved both an experimental physics group headed by Raman and a theoretical physics group headed by associate professor Carlos Sa De Melo, also in the Georgia Tech School of Physics.

The researchers first assembled a gaseous Bose-Einstein condensate that consisted of as many as three million sodium atoms cooled to nearly absolute zero. To begin the experiment, they switched on a magnetic field applied to the BEC that instantly placed the system out of equilibrium. That triggered spin-exchange collisions as the atoms attempted to transition from one ground state to a new one. Atoms near one another became entangled, pairing up with one atom’s spin pointing up, and the other’s pointing down. This pairing of opposite spins created a correlation between pairs of atoms that moved through the entire BEC as it established a new equilibrium.

The researchers, who included graduate student Anshuman Vinit and former postdoctoral fellow Eva Bookjans, measured the correlations as they spread through the cloud of cold atoms. At first, the quantum entanglement was concentrated in space, but over time, it spread outward like drop of dye diffuses through water.

“You can imagine having a drop of dye that is concentrated at one point in space,” Raman said. “Through diffusion, the dye molecules move throughout the water, slowly spreading throughout the entire system.”

The research could help scientists anticipate the operating speed for a quantum computing system composed of many cores communicating through a BEC.

“This propagation takes place on the time scale of ten to a hundred milliseconds,” Raman said. “This is the speed at which quantum information naturally flows through this kind of system. If you were to use this medium for quantum communication, that would be its natural time scale, and that would set the timing for other processes.”

Though relevant to communication of quantum information, the process also showed how a large system undergoing a phase transition does so in localized patches that expand to attempt to incorporate the entire system.

“An extended system doesn’t move from one phase to another in a uniform way,” said Raman. “It does this locally. Things happen locally that are not connected to one another initially, so you see this inhomogeneity.”

Beyond quantum computing, the results may also have implications for quantum sensing – and for the study of other physical systems that undergo phase transitions.

“Phase transitions have universal properties,” Raman noted. “You can take the phase transitions that happen in a variety of systems and find that they are described by the same physics. It is a unifying principle.”

Raman hopes the work will lead to new ways of thinking about quantum computing, regardless of its immediate practical use.

“One paradigm of quantum computing is to build a linear chain of as many trapped ions as possible and to simultaneously engineer away as many challenges as possible,” he said. “But perhaps what may be successful is to build these smaller quantum systems that can communicate with one another. It’s important to try as many things as possible and to keep an open mind. We need to try to understand these systems as well as we can.”

This research was supported by the Department of Energy (DOE) through grant DE-FG-02-03ER15450 and by the National Science Foundation under grant PHY-1100179. The conclusions in this article are those of the principal investigator and do not necessarily represent the official views of the DOE or the NSF.

CITATION: Vinit, Anshuman, et al., “Antiferromagnetic Spatial Ordering in a Quenched One-dimensional Spinor Gas, (Physical Review Letters, 2013).

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181

Media Relations Contact:  John Toon (404-894-6986)(jtoon@gatech.edu)

Writer: John Toon

None of these emails were what they pretended to be. The first directed victims to a website that asked for personal information, including the user’s password. The second included a virus launched when the “marketing plan” was opened. The third directed users to a website that attempted to install a malicious program.

All three are examples of what information security experts at the Georgia Tech Research Institute (GTRI) say is the most challenging threat facing corporate networks today: “spear phishing.”

Generic emails asking employees to open malicious attachments, provide confidential information or follow links to infected websites have been around for a long time. What’s new today is that the authors of these emails are now targeting their attacks using specific knowledge about employees and the organizations they work for. The inside knowledge used in these spear phishing attacks gains the trust of recipients.

“Spear phishing is the most popular way to get into a corporate network these days,” said Andrew Howard, a GTRI research scientist who heads up the organization’s malware unit. “Because the malware authors now have some information about the people they are sending these to, they are more likely to get a response. When they know something about you, they can dramatically increase their odds.”

The success of spear phishing attacks depends on finding the weakest link in a corporate network. That weakest link can be just one person who falls for an authentic-looking email.

“Organizations can spend millions and millions of dollars to protect their networks, but all it takes is one carefully-crafted email to let someone into it,” Howard said. “It’s very difficult to put technical controls into place to prevent humans from making a mistake. To keep these attacks out, email users have to do the right thing every single time.”

Howard and other GTRI researchers are now working to help email recipients by taking advantage of the same public information the malware authors use to con their victims. Much of that information comes from social media sites that both companies and malware authors find helpful. Other information may be found in Securities and Exchange Commission (SEC) filings, or even on an organization’s own website.

“There are lots of open sources of information that will increase the chances of eliciting a response in spear phishing,” Howard said. “We are looking at a way to warn users based on this information. We’d like to see email systems smart enough to let users know that information contained in a suspect message is from an open source and suggest they be cautious.”

Other techniques to counter the attacks may come from having access to all the traffic entering a corporate network.

To increase their chance of success, criminals attempting to access a corporate network often target more than one person in an organization. Network security tools could use information about similar spear phishing attempts to warn other members of an organization. And by having access to all email, security systems could learn what’s “normal” for each individual – and recognize unusual email that may be suspicious.

“We are looking at building behavioral patterns for users so we’d know what kinds of email they usually receive. When something comes in that’s suspicious, we could warn the user,” Howard said. “We think the real answer is to keep malicious email from ever getting into a user’s in-box, but that is a much more difficult problem.”

It’s difficult because organizations today depend on receiving, opening and responding to email from customers. Deleting or even delaying emails can have a high business cost.

“What we do requires a careful balance of protecting the user, but allowing the user to get his or her job done,” he said. “Like any security challenge we have to balance that.”

These and other strategies will be part of Phalanx, a new product being developed by GTRI researchers to protect corporate networks from spear phishing. It will be part of Titan, a dynamic framework for malicious software analysis that GTRI launched last spring.

Among the challenges ahead are developing natural language algorithms that can quickly separate potential spear phishing attacks from harmless emails. That could be done by searching for language indicating a request such as “open this attachment” or “verify your password.”

GTRI researchers been gaining experience with corporate networks based on security evaluations they’ve done, and work with GTRI’s own network – which receives millions of emails each day. Fortunately, they say, it’s not just the bad guys who are learning more.

“The chief financial officers of companies now understand the financial impacts of spear phishing, and whey they join forces with the chief information officers, there will be an urgency to address this problem,” he added. “Until then, users are the front line defense. We need every user to have a little paranoia about email.”

Research News
Georgia Institute of Technology
177 North Avenue
Atlanta, Georgia  30332-0181

Media Relations Contact: John Toon (404-894-6986)(jtoon@gatech.edu)

Writer: John Toon

The project is jointly supervised by Xiaoming Huo from ISyE and Hang Lu from ChBE, and the Institute for Data and High Performance Computing is providing seed funding for a graduate research assistant in ISyE to help advance the research. The model is being developed as a new platform that will allow for accurate phenotyping or classification of characteristics in the worms using high-throughput computing to determine the genes and pathways as well as compositions in food intake that contribute to fat accumulation.

The main objective is to develop the image processing system with pattern recognition to automatically distinguish the distinct types of lipid droplets, which are composed of fatty acid compounds, in the worms. The image analysis and classification system will systematically extract image features, efficiently learn models, and reliably predict phenotypes, or characteristics, from the images that are developed by studying the lipid droplets.

Current limitations in imaging and analysis of the lipid droplets in the worms have stunted the potential for growth, exploration, and attainable knowledge in the lipid droplet realm of research, says co-principal investigator Xiaoming Huo in ISyE.

Current methods used by the team enable them to obtain only one set of 3D images every ten seconds. A comprehensive study on the relationship between food composition and the resulting lipid analysis requires the ability to identify and classify the characteristics of hundreds of thousands of images. Researchers say that such high throughput is only manageable if the image processing and consequent prediction is automated.

The proposed research has direct applications in other problems in biology, such as neural development, stem cells, cancer diagnosis, and drug discovery. It is also potentially applicable in areas such as contemporary manufacturing of advanced nanomaterial, where a core problem is predicting the properties of produced nanomaterial.

“The research is potentially transformative because the proposed approach will develop a new technique for quantitative imaging, high-throughput experimentation, and analysis of lipid distribution and protein function in C. elegans, in pursuit of determining the unknown genetic contribution to fat storage and distribution,” says co-principal investigator Hang Lu in ChBE.

Part of the process involves microfluidics, sometimes called “Lab-on-a-Chip,” and used in the project for imaging, manipulating and sorting the animals. Combined with the statistical image analysis methods funded through the IDH seed grant, the researchers aspire to move the frontier of genetic research to the next level.

The Institute for Data and High Performance Computing is funding the seed project to enable the researchers to develop a proof-of-concept for the database system to store, analyze and describe data collected from 2D and 3D imaging of biological network structures. Researchers say the system will have the potential to fill a need for expanded data management requirements for large-scale biological data on sponsored research.

The current project will focus on two distinct biological systems for the initial datasets - plant roots and ant colonies in soil.

“These two systems share common methodological challenges and technological means of observation,” says Joshua Weitz, principal investigator and associate professor in the School of Biology. “Both plant roots and ant colonies are complex 3D networks that develop below ground and both can be observed using imaging technology in artificial and real soils.”

Weitz and co-PI Daniel Goldman, associate professor in the School of Physics, are actively researching the structure, function and dynamics of plant roots and ant colonies, respectively. They have identified that both systems lack a framework for acquiring, organizing, analyzing, visualizing and disseminating image data and identified spatial networks.

The development of an integrated database application to analyze the spatial networks in biology, which relates to an organism’s environment, is expected to enhance the pace of discovery and the reproducibility of results when the system is deployed. The researchers say that the two selected biological systems will ensure that the integrated system is truly cross-cutting in its ability to handle heterogeneous data from various sources and in various data representations.

The Georgia Tech project will facilitate key steps toward a common framework for the integrated analysis of spatial networks in biology. Weitz’s group collaborates with plant biologists to characterize root system structures of crop plants grown in “artificial soil” and of crop plants grown in natural environments. Goldman’s group will look at how the physical interaction of insects with their building materials influences nest structure.

Co-PI Alexander Bucksch, postdoctoral scientist in the School of Interactive Computing, will apply his experience on the project in analysis of large-scale networks in the plant sciences. Abhiram Das, Ph.D. candidate in Bioinformatics, is being funded through the IDH seed grant and is involved with the database development. The database application eventually will be available for other research problems.

Principal investigators Zhigang Peng, associate professor in the School of Earth and Atmospheric Sciences, and Bo Hong, assistant professor in the School of Electrical and Computer Engineering, have successfully developed a GPU-based method that can significantly accelerate the detection of earthquake signals from continuous data in a relatively small-scale space-time window (i.e. months of data recorded at several close-proximity seismic stations).

The researchers are now analyzing much larger data sets that include several years of data recorded at hundreds of seismic stations. The datasets will allow the researchers to address the challenges of scaling the current seismic-detection code to much larger real-world data and to significantly improve scientific understanding of the physics of earthquakes. The long-term goal is in developing scalable methods for seismic data analysis in the context of Big Data challenges.

“So far we have identified approximately 70 times more earthquakes around the Salton Sea geothermal field than listed in the official Southern California Seismic Network catalog,” says Peng. “These newly detected events could be used to help better understand how earthquakes are triggered in the immediate vicinity of a mainshock rupture.”

The researchers are analyzing existing recorded seismic data and taking these earlier earthquakes’ waveforms to use as templates to find “hidden” seismic activity. They automatically scan through continuous recordings from seismic stations to detect previously undetected earthquakes that have high waveform similarities to the template events.

“This is especially useful when many earthquakes occur in a short time, such as during an aftershock sequence or earthquake swarm,” said Peng. “These newly detected events are not only vital for better understanding the fundamental physics of earthquake interaction, but also useful for rapid earthquake source characterization and seismic hazard forecasting.”

The computation demands of the target problem - identifying the previously undetected earthquake signals in years of continuous data - can only be satisfied by a GPU cluster, such as Georgia Tech’s Keeneland system, due to the scale of the problem according to the researchers.

“During this new seed grant period, we plan to address the code scalability issues through innovations in code engineering and workload management,” says Hong.

The seismic detection code will be available to the scientific community and researchers hoping to generalize an understanding of the data processing challenges in the research space. This will allow researchers to extend the results to a broader range of data-intensive seismic problems such as automatic scanning and detection of new seismic events, including tremors, regular earthquakes, and glacial events.

IDH funding supports graduate research assistants Xiaofeng Meng, a Ph.D. student in Earth and Atmospheric Sciences, and Xiao Yu, a Ph.D. student in Electrical and Computer Engineering.

Papers: (9 papers of the 108, with 22.0% acceptance rate)

• Optimizing Checkpoints Using NVM as Virtual Memory. Sudarsun Kannan (Georgia Institute of Technology, USA); Ada Gavrilovska (Georgia Institute of Technology, USA); Karsten Schwan (Georgia Tech, USA); Dejan Milojicic (HP Labs, USA)
• FlexIO: I/O Middleware for Location-Flexible Scientific Data Analytics. Fang Zheng (Georgia Tech, USA); Hongbo Zou (Georgia Institute of Technology, USA); Greg Eisenhauer (Georgia Institute of Technology, USA); Karsten Schwan (Georgia Tech, USA); Matthew Wolf (Georgia Institute of Technology, USA); Jai Dayal (Georgia Institute of Technology, USA); Tuan-Anh Nguyen (Georgia Institute of Technology, USA); Jianting Cao (Beihang University, P.R. China); Mohammad Abbasi (Georgia Insitute of Technology, USA); Scott Klasky (Oak Ridge National Laboratory, USA); Norbert Podhorszki (Oak Ridge National Laboratory, USA); Hongfeng Yu (Sandia National Laborotories, USA)
• iBridge: Improving Unaligned Parallel File Access with Solid-State Drives. Xuechen Zhang (Georgia Institute of Technology, USA); Ke Liu (Wayne State University, USA); Kei Davis (Los Alamos National Laboratory, USA); Song Jiang (Wayne State University, USA)
• Energy-Efficient Scheduling for Best-Effort Interactive Services to Achieve High Response Quality. Zhihui Du (Tsinghua University, P.R. China); Hongyang Sun (Nanyang Technological University, Singapore); Yuxiong He (Microsoft Research, USA); Yu He (Tsinghua Univiversity, P.R. China); David A. Bader (Georgia Institute of Technology, USA); Huangzhe Zhang (Beijing University of Post and Telecommunication, P.R. China)
• A roofline model of energy. Jee Choi (Georgia Institute of Technology, USA); Richard W Vuduc (Georgia Institute of Technology, USA)
• A theoretical framework for algorithm-architecture co-design. Kenneth Czechowski (Georgia Institute of Technology, USA); Richard W Vuduc (Georgia Institute of Technology, USA)
• Best Paper Finalist: Extending the Generality of Molecular Dynamics Simulations on a Special-Purpose Machine. Daniele Scarpazza (D. E. Shaw Research, USA); Douglas Ierardi (D. E. Shaw Research, USA); Adam Lerer (D. E. Shaw Research, USA); Kenneth Mackenzie (D. E. Shaw Research, USA); Albert Pan (D. E. Shaw Research, USA); Joseph Bank (D. E. Shaw Research, USA); Edmond Chow (Georgia Institute of Technology, USA); Ron Dror (D. E. Shaw Research, USA); Jp Grossman (D. E. Shaw Research, USA); Daniel Killebrew (D. E. Shaw Research, USA); Mark Moraes (D. E. Shaw Research, USA); Cristian Predescu (D. E. Shaw Research, USA); John Salmon (D. E. Shaw Research, USA); David Shaw (D. E. Shaw Research, USA)
• Cura: A Cost-optimized Model for MapReduce in a Cloud. Balaji Palanisamy (Georgia Institute of Technology, USA); Aameek Singh (IBM Almaden Research Center, USA); Ling Liu (Georgia Tech, USA); Bryan Langston (IBM Research, Almaden, USA)
• Optimizing Checkpoints Using NVM as Virtual Memory. Sudarsun Kannan (Georgia Institute of Technology, USA); Ada Gavrilovska (Georgia Institute of Technology, USA); Karsten Schwan (Georgia Tech, USA); Dejan Milojicic (HP Labs, USA)

IPDPS 2013 Panel on Big Data in 10 Years

• David A. Bader, Panelist

Ph.D. Forum

• Designing Hybrid Architectures for Massive-Scale Graph Analysis. David Ediger; David A. Bader (Georgia Institute of Technology, USA)

Keynote:

• Rich Vuduc, "What first principles of algorithms and architectures says about heterogeneity." The Third International Workshop on Accelerators and Hybrid Exascale Systems (AsHES)

IPDPS 2013 Organization:

• David A. Bader, Steering Committee
• Bo Hong, PhD Forum Co-Chair
• Program Committee for IPDPS:
• David A. Bader
• Bo Hong
• Santosh Pande
• Jason Riedy
• Jeffrey Vetter
• Rich Vuduc
• David A. Bader and Srinivas Aluru, Co-Chairs, 12th IEEE International Workshop on High Performance Computational Biology (HiCOMB 2013)
• David A. Bader, Program Committee, The Third International Workshop on Accelerators and Hybrid Exascale Systems (AsHES)
• Program Committee for the Workshop on Multithreaded Architectures and Applications (MTAAP 2013):
• David A. Bader
• Bo Hong
• Jeffrey Vetter

Workshops:

• HCW 2013: 22nd International Heterogeneity in Computing Workshop
• Brawny vs. Wimpy: Evaluation and Analysis of Modern Workloads on Heterogeneous Processors. Vishal Gupta (Georgia Institute of Technology, USA), Karsten Schwan (Georgia Institute of Technology, USA)
• 12th IEEE International Workshop on High Performance Computational Biology (HiCOMB 2013)
• HPC Software Libraries for Next-Gen Sequencing Analytics. Srinivas Aluru.
• The Ninth Workshop on High-Performance, Power-Aware Computing (HPPAC)
• PowerTune: Differentiated Power Allocation in Over-provisioned Multicore Systems. Vishal Gupta and Karsten Schwan
• Workshop on Multithreaded Architectures and Applications (MTAAP 2013)
• CHiP: A Profiler to Measure the effect of Cache Contention on Scalability. Bevin Brett (Intel Corporation, USA); Pranith Kumar (Georgia Institute of Technology, USA); Minjang Kim (Georgia Institute of Technology, USA); Hyesoon Kim (Georgia Tech, USA)
• Investigating Graph Algorithms in the BSP Model on the Cray XMT. David Ediger (Georgia Institute of Technology, USA); David A. Bader (Georgia Institute of Technology, USA)
• Multithreaded Community Monitoring for Massive Streaming Graph Data. Jason Riedy (Georgia Institute of Technology, USA); David A. Bader (Georgia Institute of Technology, USA)