Quaternion optical computing chip for parallel high-dimensional data processing

TL;DR

This paper introduces the first quaternion optical computing chip (QOCC) that leverages light's parallelism to process high-dimensional data efficiently, demonstrating superior accuracy and lower computational load in various applications.

Contribution

The paper presents the design and benchmarking of a novel quaternion optical computing chip capable of high-dimensional data processing using wavelength-division multiplexing.

Findings

Higher computational fidelity (RMSE < 0.035) compared to electronic systems

2/3 reduction in computational load relative to electronic counterparts

Successful application in 3D point cloud, RGB transformation, and neural networks

Abstract

Optical computing chips have emerged as a transformative computing technology due to their high computational density, low energy consumption, and compact footprint. While real- and complex-valued computing chips have been well developed, their fundamental limitations in representing high-dimensional data significantly constrain their applicability in modern signal processing. Quaternions enable direct operations on three- and four-dimensional data, powering high-dimensional processing in data analytics and artificial intelligence. Here we demonstrate a quaternion optical computing chip (QOCC) for the first time and benchmark its performance in several typical application scenarios: three-dimensional point cloud processing, RGB chromatic transformation, and quaternion convolutional neural network for color image recognition. The QOCC harnesses high parallelism of light by wavelength-division multiplexing, processing high-dimensional data simultaneously through multiple optical wavelength channels. Compared to the electronic computing counterpart, our QOCC achieves higher computational fidelity (root mean square error < 0.035) and substantially reduced computational load (2/3 lower). It paves the way towards next-generation optical computing, overcoming the limitations of traditional computing systems in high-dimensional data processing.