PT  - JOURNAL ARTICLE
AU  - Hughes, Tyler W.
AU  - Williamson, Ian A. D.
AU  - Minkov, Momchil
AU  - Fan, Shanhui
TI  - Wave physics as an analog recurrent neural network
AID  - 10.1126/sciadv.aay6946
DP  - 2019 Dec 01
TA  - Science Advances
PG  - eaay6946
VI  - 5
IP  - 12
4099  - http://advances.sciencemag.org/content/5/12/eaay6946.short
4100  - http://advances.sciencemag.org/content/5/12/eaay6946.full
SO  - Sci Adv2019 Dec 01; 5
AB  - 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.