Analog readout for optical reservoir computers

TL;DR

This paper introduces an analog readout for optical reservoir computers that enables real-time processing, overcoming a key bottleneck and improving performance over traditional methods.

Contribution

The authors designed and experimentally tested an analog readout for optoelectronic reservoir computers, allowing real-time operation and surpassing non-reservoir methods.

Findings

Analog readout enables real-time processing in optical reservoir computers.

Performance exceeds that of non-reservoir methods on benchmark tasks.

The approach overcomes a major bottleneck in hardware reservoir computing.

Abstract

Reservoir computing is a new, powerful and flexible machine learning technique that is easily implemented in hardware. Recently, by using a time-multiplexed architecture, hardware reservoir computers have reached performance comparable to digital implementations. Operating speeds allowing for real time information operation have been reached using optoelectronic systems. At present the main performance bottleneck is the readout layer which uses slow, digital postprocessing. We have designed an analog readout suitable for time-multiplexed optoelectronic reservoir computers, capable of working in real time. The readout has been built and tested experimentally on a standard benchmark task. Its performance is better than non-reservoir methods, with ample room for further improvement. The present work thereby overcomes one of the major limitations for the future development of hardware reservoir computers.