Prof. Abhijit Chatterjee at ECE Spring 2023 Colloquium - CANCELLED
The talk has been moved to fall 2024
Error Resilient Neuromorphic Systems Using Embedded Predictive Neuron Checks
The reliability of emerging neuromorphic compute fabrics is of great concern due to their widespread use in critical data-intensive applications. Ensuring such reliability is difficult due to the intensity of underlying computations (billions of parameters), errors induced by low power operation and the complex relationship between errors in computations and their effect on network performance accuracy. In this talk, we study the problem of designing error-resilient neuromorphic systems where errors can stem from: (a) soft errors in computation of matrix-vector multiplications and neuron activations, (b) malicious trojan and adversarial security attacks and (c) effects of manufacturing process variations on analog crossbar arrays that can affect DNN accuracy. The core principle of error detection relies on embedded predictive neuron checks using invariants derived from the statistics of nominal neuron activation patterns of hidden layers of a neural network. Algorithmic encodings of hidden neuron function are also used to derive invariants for checking. A key contribution is designing checks that are robust to the inherent nonlinearity of neuron computations with minimal impact on error detection coverage. Once errors are detected, they are corrected using probabilistic methods due to the difficulties involved in exact error diagnosis in such complex systems. The technique is scalable across soft errors as well as a range of security attacks. The effects of manufacturing process variations are handled through the use of compact tests from which DNN performance can be assessed using learning techniques. Experimental results on a variety of neuromorphic test systems: DNNs, spiking networks and hyperdimensional computing are presented.
About Professor Abhijit Chatterjee
Abhijit Chatterjee is a Professor in the School of Electrical and Computer Engineering at Georgia Tech and a Fellow of the IEEE. He received his Ph.D. in electrical and computer engineering from the University of Illinois at Urbana-Champaign in 1990. Chatterjee received the NSF Research Initiation Award in 1993 and the NSF CAREER Award in 1995. He has received seven Best Paper Awards and three Best Paper Award nominations. His work on self-healing chips was featured as one of General Electric’s key technical achievements in 1992 and was cited by the Wall Street Journal. In 1995, he was named a Collaborating Partner in NASA’s New Millennium project. In 1996, he received the Outstanding Faculty for Research Award from the Georgia Tech Packaging Research Center, and in 2000, he received the Outstanding Faculty for Technology Transfer Award, also given by the Packaging Research Center. In 2007, his group received the Margarida Jacome Award for work on VIZOR: Virtually Zero Margin Adaptive RF from the Berkeley Gigascale Research Center (GSRC). Dr. Chatterjee has authored over 450 papers in refereed journals and meetings, has 22 patents and supervised over 50 Ph.D dissertations. He is a co-founder of Ardext Technologies Inc., a mixed-signal test solutions company and served as chairman and chief scientist from 2000-2002. His research interests include error-resilient machine learning, signal processing and control systems, mixed-signal/RF/multi-GHz design and test and adaptive real-time systems.
