ArchimedesSeminar Series - Towards next-generation AI with neuro-inspired computing
Dr. Angeliki Pantazi (IBM Research – Zurich)
Abstract
The recent advances in AI show that models only remotely inspired by biology can achieve impressive results on human-like tasks. These AI capabilities are achieved using the latest generations of Large Language Models (LLMs), enabled by their powerful transformer architecture. However, the operational principles of these systems have little similarity to the substantially more evolved network designs that the brain employs. Contrary to humans, the AI models operate with enormous energy and area consumption and their learning mechanism requires a large amount of data for training. A computational model inspired by the principles and mechanisms of biological systems is likely to provide more efficient innovative AI solutions due to its inherent conception idea. For example, the sparse neural activity, the neural coding strategies, the efficient neuronal dynamics, and the local plasticity mechanisms are potential key enablers for the astounding energy efficiency of the brain that can be integrated in AI systems. In this talk, I will provide an overview of our research on neuro-inspired models and learning algorithms that aim to address the sustainability concerns of current AI models. I will also showcase the efficiency of neuro-inspired computing concepts in several AI applications. Finally, I will present how neuro-inspired models could further leverage the acceleration provided by analog AI hardware.
Short Bio
Angeliki Pantazi is a Principal Research Staff Member and a Research Manager at the IBM Research – Zurich. She received her Diploma and Ph.D. degrees in Electrical Engineering and Computer Technology from the University of Patras, Greece. Since 2006, she is a Research Staff Member in IBM Research – Zurich, where she has contributed to several projects in the area of data storage systems. Currently, she leads the research on neuro-inspired computing in the Emerging Computing and Circuits group. She was named IBM Master Inventor in 2014 and became a senior member of the IEEE in 2015 and a Fellow of IFAC in 2019. She was a co-recipient of the 2009 IEEE Control Systems Technology Award, the 2009 IEEE Transactions on Control Systems Technology Outstanding Paper Award and the 2014 IFAC Industrial Achievement Award. In 2017, she received an IBM Corporate Award and the IEEE Control Systems Society Transition to Practice Award. She has published over 100 refereed articles and holds over 50 granted patents.
________________________________________________________________________________
Meeting ID: 362 849 199 025
Passcode: 6piKyG
________________________________________________________________________________
