Integrated photonic neural networks provide a promising platform for energy-efficient, high-throughput machine learning with extensive scientific and commercial applications. Photonic neural networks efficiently transform optically encoded inputs using Mach-Zehnder interferometer mesh networks interleaved with nonlinearities. We experimentally trained a three-layer, four-port silicon photonic neural network with programmable phase shifters and optical power monitoring to solve classification tasks using “in situ backpropagation,” a photonic analog of the most popular method to train conventional neural networks. We measured backpropagated gradients for phase-shifter voltages by interfering forward- and backward-propagating light and simulated in situ backpropagation for 64-port photonic neural networks trained on MNIST image recognition given errors. All experiments performed comparably to digital simulations ( > 94% test accuracy), and energy scaling analysis indicated a route to scalable machine learning.

Miniaturizing particle accelerators: Particle accelerators are usually associated with large national facilities. Because photons are able to impart momentum to electrons, there are also efforts to develop laser-based particle accelerators. Sapra et al. developed an integrated particle accelerator using photonic inverse design methods to optimize the interaction between the light and the electrons. They show that an additional kick of around 0.9 kilo–electron volts (keV) can be given to a bunch of 80-keV electrons along just 30 micrometers of a specially designed channel. Such miniaturized dielectric laser accelerators could open up particle physics to a number of scientific disciplines. Science , this issue p. 79

Complex optical devices including aspherical focusing mirrors, solar concentrator arrays, and immersion lenses were 3D printed using commercial technology and experimentally demonstrated by evaluating surface roughness and shape. The as-printed surfaces had surface roughness on the order of tens of microns. To improve this unacceptable surface quality for creating optics, a polymer smoothing technique was developed. Atomic force microscopy and optical profilometry showed that the smoothing technique reduced the surface roughness to a few nanometers, consistent with the requirements of high-quality optics, while tests of optical functionality demonstrated that the overall shapes were maintained so that near theoretically predicted operation was achieved. The optical surface smoothing technique is a promising approach towards using 3D printing as a flexible tool for prototyping and fabrication of miniaturized high-quality optics.

Reconfigurable photonic mesh networks of tunable beamsplitter nodes can linearly transform N-dimensional vectors representing input modal amplitudes of light for applications such as energy-efficient machine learning hardware, quantum information processing, and mode demultiplexing. Such photonic meshes are typically programmed and/or calibrated by tuning or characterizing each beam splitter one-by-one, which can be time-consuming and can limit scaling to larger meshes. Here we introduce a graph-topological approach that defines the general class of feedforward networks and identifies columns of non-interacting nodes that can be adjusted simultaneously. By virtue of this approach, we can calculate the necessary input vectors to program entire columns of nodes in parallel by simultaneously nullifying the power in one output of each node via optoelectronic feedback onto adjustable phase shifters or couplers. This parallel nullification approach is fault-tolerant to fabrication errors, requiring no prior knowledge or calibration of node parameters and reducing programming time by a factor of order N to being proportional to the optical depth (number of node columns). As a demonstration, we simulate our programming protocol on a feedforward optical neural network model trained to classify handwritten digit images from the MNIST dataset with up to 98% validation accuracy.

Single-photon light detection and ranging (LiDAR) techniques use emerging single-photon detectors (SPADs) to push 3D imaging capabilities to unprecedented ranges. However, it remains challenging to robustly estimate scene depth from the noisy and otherwise corrupted measurements recorded by a SPAD. Here, we propose a deep sensor fusion strategy that combines corrupted SPAD data and a conventional 2D image to estimate the depth of a scene. Our primary contribution is a neural network architecture—SPADnet—that uses a monocular depth estimation algorithm together with a SPAD denoising and sensor fusion strategy. This architecture, together with several techniques in network training, achieves state-of-the-art results for RGB-SPAD fusion with simulated and captured data. Moreover, SPADnet is more computationally efficient than previous RGB-SPAD fusion networks.

As blockchain technology and cryptocurrency become increasingly mainstream, photonic computing has emerged as an efficient hardware platform that reduces ever-increasing energy costs required to verify transactions in decentralized cryptonetworks. To reduce sensitivity of these verifications to photonic hardware error, we propose and experimentally demonstrate a cryptographic scheme, LightHash, that implements robust, low-bit precision matrix multiplication in programmable silicon photonic networks. We demonstrate an error mitigation scheme to reduce error by averaging computation across circuits, and simulate energy-efficiency-error trade-offs for large circuit sizes. We conclude that our error-resistant and efficient hardware solution can potentially generate a new market for decentralized photonic blockchain.