Photonic Tensor Core
Photonic tensor cores (PTCs) are optical computing components designed to accelerate artificial intelligence (AI) computations, primarily matrix multiplications crucial for deep neural networks. Current research focuses on optimizing PTC design through automated topology search algorithms, algorithm-hardware co-design for sparsity exploitation, and time-multiplexing techniques to improve energy efficiency and reduce hardware footprint. These advancements aim to overcome limitations in reconfigurability, power consumption, and precision, ultimately enabling more efficient and powerful AI accelerators for both edge and cloud computing applications.
Papers
Integrated multi-operand optical neurons for scalable and hardware-efficient deep learning
Chenghao Feng, Jiaqi Gu, Hanqing Zhu, Rongxing Tang, Shupeng Ning, May Hlaing, Jason Midkiff, Sourabh Jain, David Z. Pan, Ray T. Chen
M3ICRO: Machine Learning-Enabled Compact Photonic Tensor Core based on PRogrammable Multi-Operand Multimode Interference
Jiaqi Gu, Hanqing Zhu, Chenghao Feng, Zixuan Jiang, Ray T. Chen, David Z. Pan