A small microring array that performs large complex-valued matrix-vector multiplication
큰 복소수 행렬-벡터 곱셈을 수행하는 작은 마이크로링 어레이
Una pequeña matriz de microanillos que realiza una gran multiplicación de matriz-vector de valor complejo
Un petit réseau de micro-anneaux qui effectue une grande multiplication matrice-vecteur à valeurs complexes
Небольшой массив микроколец, который выполняет большое комплексное умножение матрицы на вектор
Junwei Cheng 成骏伟 ¹, Yuhe Zhao 赵雨赫 ¹, Wenkai Zhang ¹, Hailong Zhou 周海龙 ¹ ² ³, Dongmei Huang 黄冬梅 ³ ⁴, Qing Zhu ⁵, Yuhao Guo 郭宇昊 ⁵, Bo Xu ⁵, Jianji Dong 董建绩 ¹, Xinliang Zhang 张新亮 ¹
¹ Wuhan National Laboratory for Optoelectronics, School of Optical and Electronic Information, Huazhong University of Science and Technology, Wuhan 430074, China
中国 武汉 华中科技大学光学与电子信息学院 武汉光电国家实验室
² Photonics Research Centre, Department of Electronic and Information Engineering, The Hong Kong Polytechnic University, Hong Kong 999077, China
中国 香港 香港理工大学 电子及资讯工程学系 光子技术研究院
³ The Hong Kong Polytechnic University Shenzhen Research Institute, Shenzhen 518057, China
中国 深圳 香港理工大学深圳研究院
⁴ Photonics Research Centre, Department of Electrical Engineering, The Hong Kong Polytechnic University, Hong Kong 999077, China
中国 香港 香港理工大学 电机工程学系 光子技术研究院
⁵ Institute of Strategic Research, Huawei Technologies, Shenzhen 518129, China
中国 深圳 华为战略研究院
Frontiers of Optoelectronics, 15 Jun 2022

As an important computing operation, photonic matrix–vector multiplication is widely used in photonic neutral networks and signal processing. However, conventional incoherent matrix–vector multiplication focuses on real-valued operations, which cannot work well in complex-valued neural networks and discrete Fourier transform.

In this paper, we propose a systematic solution to extend the matrix computation of microring arrays from the real-valued field to the complex-valued field, and from small-scale (i.e., 4 × 4) to large-scale matrix computation (i.e., 16 × 16). Combining matrix decomposition and matrix partition, our photonic complex matrix–vector multiplier chip can support arbitrary large-scale and complex-valued matrix computation. We further demonstrate Walsh-Hardmard transform, discrete cosine transform, discrete Fourier transform, and image convolutional processing.

Our scheme provides a path towards breaking the limits of complex-valued computing accelerator in conventional incoherent optical architecture. More importantly, our results reveal that an integrated photonic platform is of huge potential for large-scale, complex-valued, artificial intelligence computing and signal processing.
Frontiers of Optoelectronics_1
Frontiers of Optoelectronics_2
Frontiers of Optoelectronics_3
Frontiers of Optoelectronics_4
Reviews and Discussions
Robust measurement of orbital angular momentum of a partially coherent vortex beam under amplitude and phase perturbations
Deblurring, artifact-free optical coherence tomography with deconvolution-random phase modulation
Dynamic interactive bitwise meta-holography with ultra-high computational and display frame rates
Multi-dimensional multiplexing optical secret sharing framework with cascaded liquid crystal holograms
Physics-informed deep learning for fringe pattern analysis
Advancing computer-generated holographic display thanks to diffraction model-driven deep nets
Inverse design for material anisotropy and its application for a compact X-cut TFLN on-chip wavelength demultiplexer
Improved spatiotemporal resolution of anti-scattering super-resolution label-free microscopy via synthetic wave 3D metalens imaging
Wide-spectrum optical synthetic aperture imaging via spatial intensity interferometry
Flat soliton microcomb source
Smart palm-size optofluidic hematology analyzer for automated imaging-based leukocyte concentration detection
Deep learning enabled single-shot absolute phase recovery in high-speed composite fringe pattern profilometry of separated objects

Previous Article                                Next Article
Copyright © Hot Paper