High-speed optical information processing on chips inspired by human brain

April 1, 2014
Small neural network

16-node optical neural network (credit: Ghent University)

Ghent University researchers have created a small 16-nodes neural network in a silicon photonics chip, inspired by how the human brain works.

The goal is to create a new information technique based on light instead of electricity, with the potential for high speed (up to several hundreds of Gbits/sec., or more with miniaturization), low power consumption, and compact design.

The researchers have experimentally shown that the chip can be used for a diverse range of tasks, such as Boolean logic operations, basic machine-intelligence tasks (classification and regression), and limited speech recognition of spoken digits.

This study, described in Nature Communications (open access), was funded by the European Research Council (ERC) Starting Grant from the Naresco and by the Belgian IAP program via the [email protected] network.

Abstract of Nature Communications paper

In today’s age, companies employ machine learning to extract information from large quantities of data. One of those techniques, reservoir computing (RC), is a decade old and has achieved state-of-the-art performance for processing sequential data. Dedicated hardware realizations of RC could enable speed gains and power savings. Here we propose the first integrated passive silicon photonics reservoir. We demonstrate experimentally and through simulations that, thanks to the RC paradigm, this generic chip can be used to perform arbitrary Boolean logic operations with memory as well as 5-bit header recognition up to 12.5 Gbit s−1, without power consumption in the reservoir. It can also perform isolated spoken digit recognition. Our realization exploits optical phase for computing. It is scalable to larger networks and much higher bitrates, up to speeds >100 Gbit s−1. These results pave the way for the application of integrated photonic RC for a wide range of applications.