Virginia Tech® home

Bio-Semiconductor Neural Interface: Towards Bridging Natural Intelligence and AI

Wang

We are honored to host Dr. Jun Wang, Assistant Professor of Electrical and Computer Engineering at Virginia Tech, whose work advances high-density neural recording and low-power neuromorphic chip systems. His background includes a Ph.D. in Bioengineering from UC San Diego and postdoctoral research at Harvard University.

Abstract
How closely should an artificial neural network (ANN) emulate a biological neuronal network (BNN), then it will begin to exhibit the same type of intelligence of BNN? To tackle this profound question, I initiated my exploration with the question: “Is it possible to substitute a portion of biological neurons with artificial ones without altering the pre-existing neuronal functionality?”

Developing a cyto-neuromorphic hybrid system provides a promising path toward answering this question. In this talk, I will present our ongoing efforts toward such a system through bio-semiconductor neural interfaces:(1) Advancing synaptic connectivity mapping from in-vitro, ex-vivo, to in-vivo, enabled by high-throughput neural interface system-on-chip designs.
(2) Enhancing neuromorphic system designs using synaptic connectivity maps, demonstrated by using memristors to extract synaptic connection strength and CMOS integrated circuits to faithfully emulate neuronal dynamics.
(3) Investigating bio-semiconductor hybrid systems at the core of neuromorphic systems, exemplified by a prototype of a cyto-silicon hybrid system with closed-loop modulation.

Biography
Jun Wang is an assistant professor of Department of Electrical and Computer Engineering. Before joining Virginia Tech, he was a postdoctoral research associate at Harvard. He obtained his Ph.D. in Bioengineering from UC San Diego in 2019. His current research interests include high-density CMOS-nano system-on-chip design for high-resolution neural recording and modulation, low-power mixed-signal neuromorphic chip design, and bio-semiconductor hybrid system implementation.

By Pouya Faeghi