Mohsen Imani

UC Irvine

Hyperdimensional Computing for Efficient, Robust and Interpretable Cognitive Learning

Abstract

With increasing demand for efficient processing of diverse cognitive tasks using vast amounts of data, intelligent devices are becoming increasingly necessary. This talk explores strategies for accelerating algorithms in hardware while also redesigning them to more closely model the human brain. We introduce Hyperdimensional Computing (HDC) as an alternative computing method that mimics important brain functionalities for high-efficiency and noise-tolerant computation. HDC is an end-to-end framework that enables adaptive, robust, efficient, and transparent cognitive learning by using vector operations to mimic important functionalities of human memory. A key advantage of HD computing is its ability to train on one or few shots, which means data is learned from few examples in a single pass over the training data, as opposed to many iterations. Additionally, HD computing is highly tolerant of errors, as it operates over random hypervectors that are independent and identically distributed. This makes HDC a promising solution for today’s embedded devices with limited resources and future computing systems with high noise and variability issues. This talk will showcase HDC’s capability in enabling efficient, robust, and human-interpretable learning and reasoning. Furthermore, HDC’s robustness will be leveraged to design the next generation of cognitive platforms that are highly approximate, parallel, and efficient, inspired by the human brain’s effective handling of noisy neurons.

Biography

Mohsen Imani is an Assistant Professor in the Department of Computer Science at UC Irvine. He is also a director of Bio-Inspired Architecture and Systems Laboratory (BIASLab). He is working on a wide range of practical problems in the area of brain-inspired computing, machine learning, computer architecture, and embedded systems. His research goal is to design real-time, robust, and programmable computing platforms that can natively support a wide range of learning and cognitive tasks on edge devices. Dr. Imani received his Ph.D. from the Department of Computer Science and Engineering at the UC San Diego. He has a stellar record of publication with over 170 papers in top conferences/journals. His research was also the main initiative in opening up multiple industrial and governmental research programs on brain-inspired hyperdimensional computing. Dr. Imani research has been recognized with several awards, including DARPA Young Faculty Award, DARPA Riser Award, the Bernard and Sophia Gordon Engineering Leadership Award, the Best Doctorate Research from UCSD, and several best paper awards and nomination at multiple top conferences including Design Automation Conference (DAC) in 2019 and 2020, Design Automation and Test in Europe (DATE) in 2020 and 2022, and International Conference on Computer-Aided Design (ICCAD) in 2020.