Optical computing

Silicon photonics can be used to create high-speed large-scale neuromorphic systems for artificial intelligent tasks. Here, the authors discuss the design details and behavior of a resonator spiking neuron that can be fabricated in a commercial silicon photonics foundry process.

Yibo Dong et al. implement a compact and robust diffractive neural network chip with a virtually unlimited lifetime for optical inference. The chip demonstrates high accuracy and high stability even after high temperature aging, aiming at applications in extreme environments.

Mochou Yang and colleagues describe a a ghost imaging framework based on laser mode speckle pattern which improves imaging quality at low sampling rate. The feasibility of this method is demonstrated in a turbid water environment.

Photonic Ising machines exploit the parallelism and high propagation speed of light to solve combinatorial optimization tasks. The authors propose and demonstrate a photonic Ising machine with a fully reconfigurable optical vector-matrix transformation system and a modified algorithm based on simulated annealing, solving 20 and 30-spin Ising problems with high ground state probability.

Optical techniques adopted in optical computing rely on spatial multiplexing, requiring numerous integrated elements and restricting the architecture to perform a single kernel convolution per layer. The authors demonstrate a fiber-optic computing architecture based on temporal multiplexing that performs multiple convolutions in a single layer.

Signal processing is key to communications and video image processing for astronomy, medical diagnosis, autonomous driving, big data and AI. Menxi Tan and colleagues report a photonic processor operating at 17Tb/s for ultrafast robotic vision and machine learning.

Bablich and colleagues report an Intrinsic Photomixing Detector (IPD), a cost-effective amorphous silicon device, enabling high-bandwidth, long-distance, and high accuracy Time-of-Flight optical ranging even in low light conditions. The device allows for large-scale integration on silicon or flexible platforms.

Jiashuo Shi and colleagues build an integrated camera capable of tracking objects of interest. They use optical computing to arrange molecules in the liquid crystal mask for enhanced distinction between the object and background.

Stanley Cheung and co-authors introduce co-integrated III-V/Si memristors with fundamental photonic building blocks used in both communication and computing applications. This allows a path towards realizing low-loss, non-volatile optical elements with near-zero static power consumption.

Optical beams carrying orbital angular momentum (OAM) are promising candidates for free-space optical communication. The authors devise a hybrid optical-electronic convolutional neural network approach reaching a 4-bit OAM-coded signal demultiplexing accuracy of 72.84% under strong atmospheric turbulence conditions with 3.2 times faster training time than all electronic convolutional neural network.

On-chip photonic neural network processors speed up image processing, but their scalability is limited by the number of input/output channels. Here, the authors develop a scalable image processing approach based on the compressive acquisition of real-world visual information via a single-input channel, and apply their setup as a compressive temporal encoder for highspeed imaging.

Daniel Perez Lopez is Co-founder and Chief Technology Officer at iPronics, a company dedicated to the development and commercialization of integrated programmable photonic circuits. His company focuses both on hardware advances for novel circuit and component architectures as well as software advances leading to the creation of fault-tolerant automated routines enabling advanced optical networking and processing, specially for AI infrastructure and intra-datacenter communications. As a young entrepreneur, Daniel shares with us his experiences and insights of the academia-industry transition and building a spin-out company from his university research.