The award-winning poster, titled “Towards Bit Error Robust Energy-Efficient Autonomous Systems”, presents a cross-layer robust learning framework aimed at achieving aggressive energy-saving yet computational-resilient autonomous swarms.

The research tackles the challenges of the distributed resource-constrained nodes, complex cyber-physical autonomous systems and increased chip bit failures, by proposing innovative approaches including robust offline/on-device learning, collaborative swarm optimizations, and dynamic thermal-payload adjustments.

Applied across hardware chips, swarm systems, and various autonomous scenarios, the technique demonstrates consistent performance-efficiency-robustness improvements for multi-drone systems.

This acclaimed research is a product of a collaborative effort between Georgia Tech, IBM Research, Harvard University, and Lawrence Livermore National Laboratory (LLNL). The work is supported in part by the Center for the Co-Design of Cognitive Systems (CoCoSys), the IAPRA MicroE4AI program, and the Center for Research into Novel Compute Hierarchies (CRNCH) PhD Fellowship.

Zishen is advised by ECE professors Arijit Raychowdhury and Tushar Krishna. His research interests center around computer architecture, VLSI, and embedded systems, with a focus on building hardware and systems for cognitive intelligence and autonomous machines, aiming to advance their performance, efficiency, resilience, and trustworthiness.