Brain-inspired computing for intelligent autonmous systems focus of new center

UNIVERSITY PARK, Pa. — Vijay Narayanan, A. Robert Noll Chair in Engineering and distinguished professor of computer science and engineering in the School of Electrical Engineering and Computer Science, and John Sustersic, assistant research professor at the Applied Research Laboratory, Penn State, are part of a new national center that aims to develop brain-inspired computing for intelligent autonomous systems such as drones and personal robots capable of operating without human intervention.

The Center for Brain-inspired Computing Enabling Autonomous Intelligence, or C-BRIC, is a five-year project supported by $27 million in funding from the Semiconductor Research Corporation via its Joint University Microelectronics Program, which provides funding from a consortium of industrial sponsors as well as from the Defense Advanced Research Projects Agency. C-BRIC is one of six centers chosen across the nation to undertake high-risk, high-payoff research that addresses existing and emerging challenges in microelectronic technologies.

Narayanan will lead the Distributed Intelligence Theme for the center, exploring a shift from the predominant data-center-oriented machine learning and inference used today to more distributed, mobile-friendly intelligence. The technologies are expected to bring more intelligence and autonomy to edge devices such as sensors, phones and drones.

“Intelligence of most current mobile devices and sensors relies on transmitting the data and computing in a remote server. This consumes significant energy in transmitting the data and increases latency in decision-making. Further, if there is disruption to the connection, you have no decision-making ability,” explained Narayanan. “Imagine your language-translating app working when traveling abroad without having to pay roaming charges for data, connecting to a server or running out of your battery quickly. The distributed intelligence research will explore new learning and inference approaches that rely on power-efficient, battery-friendly machine learning approaches on a mobile device.”

The research will also build larger intelligent systems composed of a group of collaborating small devices. For example, analyzing a scene from viewpoints of multiple cameras can help understand the scene better. Such technology can aid robust drone navigation or provide more accurate visual assistance for the visually impaired.

Sustersic’s research within C-BRIC will focus on using sparse reinforcement and constraints to enable autonomous systems to self-guide their learning.

“Most machine learning/AI systems today require extensive training data that is expensive to collect and ultimately may or may not be ideal for practical purposes,” he said. “We are interested in how to enable learning with task-oriented constraints that are understandable and meaningful to human collaborators, to allow autonomous systems to learn with sparse human interaction and without significant quantities of expensive, labeled data.”

C-BRIC, led by Purdue University, also involves researchers from Arizona State University, Georgia Institute of Technology, Portland State University, Princeton University, University of Pennsylvania, University of Southern California and University of Illinois Urbana–Champaign. For more information on C-BRIC, as well as the other university research centers, visit  

Last Updated July 02, 2018