Neural Networks Based on Ultrafast Time-Delayed Effects in Exciton Polaritons
- 23 May 2022
- journal article
- research article
- Published by American Physical Society (APS) in Physical Review Applied
- Vol. 17 (5) , 054037
Abstract
We demonstrate that time-delayed nonlinear effects in exciton polaritons can be used to construct neural networks where information is coded in optical pulses arriving consecutively on the sample. The highly nonlinear effects are induced by time-dependent interactions with the excitonic reservoir. These nonlinearities allow the creation of a nonlinear XOR logic gate that can perform operations on the picosecond timescale. An optoelectronic neural network based on the constructed logic gate performs the classification of spoken digits with a high accuracy rate.Keywords
Funding Information
- National Science Center, Poland (2019/33/N/ST3/02019, 2017/27/B/ST3/00271, 2015/18/E/ST3/00558, 2020/37/B/ST3/01657, 2020/04/X/ST7/01379, 2019/35/N/ST3/01379, 2020/37/B/ST3/01657)
This publication has 34 references indexed in Scilit:
- Single-chip microprocessor that communicates directly using lightNature, 2015
- Deep learningNature, 2015
- Perceptrons with Hebbian Learning Based on Wave Ensembles in Spatially Patterned PotentialsPhysical Review Letters, 2015
- Broadcast and Weight: An Integrated Network For Scalable Photonic Spike ProcessingJournal of Lightwave Technology, 2014
- All-optical polariton transistorNature Communications, 2013
- Quantum fluids of lightReviews of Modern Physics, 2013
- Parallel photonic information processing at gigabyte per second data rates using transient statesNature Communications, 2013
- Polariton condensate transistor switchPhysical Review B, 2012
- Information processing using a single dynamical node as complex systemNature Communications, 2011
- Extreme learning machine: Theory and applicationsNeurocomputing, 2006