Title: Advances in Neuromorphic Computing for Fast, Efficient, and Intelligent Processing
Speaker: Mike Davies, Intel Labs
Abstract: The past fifty years have brought enormous progress in computer architecture, semiconductor scaling, and artificial intelligence, yet our computing technology today still lags far behind biological brains in many respects. While deep artificial neural networks have provided breakthroughs in AI, these gains come with heavy compute and data requirements relative to their biological counterparts. Neuromorphic computing aims to narrow this gap by drawing inspiration from the form and function of biological neural circuits, re-thinking computing from transistors to software informed by biological principles. The pace of neuromorphic research has accelerated in recent years, with chips like Intel’s Loihi providing, for the first time, compelling quantitative results over a range of workloads—from sensory perception to data efficient learning to closed-loop adaptive control to combinatorial optimization. This progress suggests the approaching commercial viability of a new class of chips that can autonomously process complex data streams, adapt, plan, behave, and learn in real time at extremely low power levels. This talk surveys some of these recent developments and remaining challenges facing the field.
Bio: Mike Davies is a Senior Principal Engineer in Intel Labs and the Director of Intel’s Neuromorphic Computing Lab. Since 2014 he has been researching neuromorphic architectures, algorithms, software, and systems, and has fabricated several neuromorphic chip prototypes, including the Loihi series. He was a founding employee of Fulcrum Microsystems and Director of its silicon engineering group until Intel’s acquisition of Fulcrum in 2011. He led the development of four generations of low latency, highly integrated Ethernet switches using Fulcrum’s proprietary asynchronous design methodology. He received B.S. and M.S. degrees from Caltech in 1998 and 2000, respectively.