JUNCA Talk "Bridging Learning and Reasoning: A Cross-Layer Approach for Neuro-Symbolic AI"

Prof. Arijit Raychowdhury, Georgia Institute of Technology, will give talk on „Bridging Learning and Reasoning: A Cross-Layer Approach for Neuro-Symbolic AI“.

You may also join the hybrid talk online via this link.

Abstract:

Neuro-symbolic AI presents an emerging compositional paradigm that fuses neural learning with symbolic reasoning to enhance the transparency, interpretability, and trustworthiness of AI solutions for complex real-world tasks. However, achieving real-time, energy-efficient, and scalable deployment requires cross-disciplinary integration through application discovery, systems thinking, and co-design across the stack.

This talk will present our recent work on a vertically-integrated approach to enabling efficient and scalable neuro-symbolic computing, from workload benchmarking, hardware architecture, to FPGA prototyping and silicon engineering. I will present workload characterization to uncover key system behavior of neuro-symbolic models. Second, we will introduce CogSys, a reconfigurable hardware architecture with streaming dataflow for real-time neuro-symbolic computing. Then, we will present NSFlow, an end-to-end FPGA framework with automated architecture generator for agile neuro-symbolic deployment. Finally, a 40nm programmable heterogeneous SoC with hybrid RRAM/SRAM based memory-centric computing for neuro-symbolic AI models will be presented. Through this synergistic cross-layer co-design, I will describe the feasibility of deploying efficient neuro-symbolic cognitive systems at scale.

Arijit Raychowdhury is the Steve W Chaddick Chair and Professor of the School of Electrical and Computer Engineering at the Georgia Institute of Technology. He joined Georgia Tech in January 2013. From 2013 to July 2019, he was an Associate Professor and held the ON Semiconductor Junior Professorship in the department. From 2020 to 2021, he held the Motorola Professorship in the ECE. Prior to joining Georgia Tech, he held research positions at Intel Corporation for six years and at Texas Instruments for one and a half years. He received his Ph.D. degree in Electrical and Computer Engineering from Purdue University in 2007. His research interests include low power digital and mixed-signal circuit design, design of power converters, signal-processors, and exploring interactions of circuits with device technologies. Dr. Raychowdhury holds more than 27 U.S. and international patents and has published over 400 articles in journals and refereed conferences. He is the winner of several prestigious awards, including the SRC Technical Excellence Award 2021, Qualcomm Faculty Awards in 2021, 2020, IEEE/ACM Innovator under 40 Award, the NSF CISE Research Initiation Initiative Award (CRII) 2015, Intel Labs Technical Contribution Award 2011, Dimitris N. Chorafas Award for outstanding doctoral research and best thesis 2007, and several fellowships. He and his students have won 18 best paper awards over the years. Dr. Raychowdhury is a Fellow of the IEEE.