Photonics reservoir computing
Project Leader: Ampalavanapillai Nirmalathas
Staff: Elaine Wong, Ampalavanapillai Nirmalathas
Collaborators: Professor Ming Wu (UC Berkeley USA) Professor Jose Capmany (UPV Spain) Dr Linh Nguyen (DST) Dr Manik Attygalle (DST)
Primary Contact: Ampalavanapillai Nirmalathas (email@example.com)
Keywords: artificial intelligence; photonics and electronics
Disciplines: Electrical & Electronic Engineering
Many applications in Radars, Lidars, Cognitive Communications and Networking and Highspeed Trading, new superfast, energy efficient, compact, and easy-to-integrate computing architectures need to be developed to create the ability for autonomous processing of spatiotemporal data streams using machine learning. As a response to this challenge, photonic reservoir computing approaches have attracted a great deal of research interest as they rely on a fixed network of neurons (‘reservoir’) implemented in the photonic domain and a fixed set of mapping of inputs to the reservoir. By simply combining the outputs of this fixed reservoir via a liner weighted combination, machine learning with much simpler learning approaches can become very effective in a fast and elegant manner. This project seeks to develop new architectures for photonic implementations of reservoir computing and also investigate their effectiveness in solving challenging computational tasks in radars, lidars, networking and imaging applications.