UC Davis is leading a new 5-year project titled Expedition on Brain-Derived Neuromorphic Computing with Intelligent Photonic and Electronic Materials (ExPlor) funded by the Air Force Office of Scientific Research, or AFOSR. Successful progress of this project will lead to $4.75 million funding to support the algorithms and architectural studies at UC Davis (S. J. Ben Yoo– Lead PI, Daniel Cox, Rishidev Chaudhuri, Randall OReilly), UC Berkeley (Kristof Bouchard), Stanford (H.-S. Philip Wong, David A.B.Miller, Alberto Salleo), MIT (Lionel Kimerling, Juejun Hu), and George Washington University (Volker Sorger).
The ExPlor project investigates brain-derived biophysical principles and algorithms applied in a hierarchical network of spiking optoelectronic (macro-circuits) and electronic (micro-circuits) neural networks, including emerging concepts in nanoscale dendrite computing (nano-circuits), synaptic stochasticity, and self-learning. Through the proposed efforts, the ExPlor project aims at understanding and realizing a biologically plausible neuromorphic system that is hierarchical, scalable, intelligent, efficient, and high-throughput.
S. J. Ben Yoo, Distinguished Professor of Electrical and Computer Engineering at UC Davis will be leading this new project. While the algorithms and software behind Artificial Intelligence and Machine Learning (AI and ML) have made impressive progress in the past decades, the limited progress in hardware architecture and technologies is causing exponential increases in power consumption in data centers which are already consuming megawatts of power, Yoo said. This new project pursues innovations in brain-like neuromorphic computing architectures realized by the integration of nanoscale photonics and electronics towards achieving million to billion times improvements in power efficiency and throughput compared to the state of the art.
