An optoacoustic field-programmable perceptron for recurrent neural networks

Abstract

A critical feature in signal processing is the ability to interpret correlations in time series signals, such as speech. Machine learning systems process this contextual information by tracking internal states in recurrent neural networks (RNNs), but these can cause memory and processor bottlenecks in applications from edge devices to data centers, motivating research into new analog inference architectures. But whereas photonic accelerators, in particular, have demonstrated big leaps in uni-directional feedforward deep neural network (DNN) inference, the bi-directional architecture of RNNs presents a unique challenge: the need for a short-term memory that (i) programmably transforms optical waveforms with phase coherence , (ii) minimizes added noise, and (iii) enables programmable readily scales to large neuron counts. Here, we address this challenge by introducing an optoacoustic recurrent operator (OREO) that simultaneously meets (i,ii,iii). Specifically, we experimentally demonstrate an OREO that contextualizes and computes the information carried by a sequence of optical pulses via acoustic waves. We show that the acoustic waves act as a link between the different optical pulses, capturing the optical information and using it to manipulate the subsequent operations. Our approach can be controlled completely optically on a pulse-by-pulse basis, offering simple reconfigurability for a use case-specific optimization. We use this feature to demonstrate a recurrent drop-out, which excludes optical input pulses from the recurrent operation. We furthermore apply OREO as an acceptor to recognize up-to $27$ patterns in a sequence of optical pulses. Finally, we introduce a DNN architecture that uses the OREO as bi-directional perceptrons to enable new classes of DNNs in coherent optical signal processing.

References in corpus (23)

• Deep Learning with Coherent Nanophotonic Circuits
• All-Optical Machine Learning Using Diffractive Deep Neural Networks
• Parallel convolution processing using an integrated photonic tensor core
• Photonics for artificial intelligence and neuromorphic computing
• Neuromorphic Silicon Photonic Networks
• Single-chip photonic deep neural network for instantaneous image classification
• All Optical Neural Network with Nonlinear Activation Functions
• Reinforcement Learning in a large scale photonic Recurrent Neural Network
• A photonic integrated circuit based erbium-doped amplifier
• Wave Physics as an Analog Recurrent Neural Network
• Analysis of Diffractive Optical Neural Networks and Their Integration with Electronic Neural Networks
• Breaking time-reversal symmetry with acoustic pumping of nanophotonic circuits
• Electrically-driven Acousto-optics and Broadband Non-reciprocity in Silicon Photonics
• Nonreciprocal Inter-band Brillouin Modulation
• Delocalized Photonic Deep Learning on the Internet's Edge
• Deep Learning with Coherent VCSEL Neural Networks
• Scalable Optical Learning Operator
• Tutorial: Photonic Neural Networks in Delay Systems
• A chip-integrated coherent photonic-phononic memory
• Observation of stimulated Brillouin scattering in silicon nitride integrated waveguides
• Arbitrary linear transformations for photons in the frequency synthetic dimension
• A Hamiltonian treatment of stimulated Brillouin scattering in nanoscale integrated waveguides
• On-chip multi-stage optical delay based on cascaded Brillouin light storage

Cited by in corpus (3)

• High-speed coherent photonic random-access memory in long-lasting sound waves
• Brillouin-based storage of QPSK signals with fully tunable phase retrieval
• Giant Brillouin gain in frozen CS2 capillaries