Wave physics as an analog recurrent neural network.

Autor:	Hughes TW; Department of Applied Physics, Stanford University, Stanford, CA 94305, USA., Williamson IAD; Department of Electrical Engineering, Stanford University, Stanford, CA 94305, USA., Minkov M; Department of Electrical Engineering, Stanford University, Stanford, CA 94305, USA., Fan S; Department of Electrical Engineering, Stanford University, Stanford, CA 94305, USA.
Jazyk:	angličtina
Zdroj:	Science advances [Sci Adv] 2019 Dec 20; Vol. 5 (12), pp. eaay6946. Date of Electronic Publication: 2019 Dec 20 (Print Publication: 2019).
DOI:	10.1126/sciadv.aay6946
Abstrakt:	Analog machine learning hardware platforms promise to be faster and more energy efficient than their digital counterparts. Wave physics, as found in acoustics and optics, is a natural candidate for building analog processors for time-varying signals. Here, we identify a mapping between the dynamics of wave physics and the computation in recurrent neural networks. This mapping indicates that physical wave systems can be trained to learn complex features in temporal data, using standard training techniques for neural networks. As a demonstration, we show that an inverse-designed inhomogeneous medium can perform vowel classification on raw audio signals as their waveforms scatter and propagate through it, achieving performance comparable to a standard digital implementation of a recurrent neural network. These findings pave the way for a new class of analog machine learning platforms, capable of fast and efficient processing of information in its native domain. (Copyright © 2019 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC).)
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu