Lin Zhongjin, Shastri Bhavin J, Yu Shangxuan, Song Jingxiang, Zhu Yuntao, Safarnejadian Arman, Cai Wangning, Lin Yanmei, Ke Wei, Hammood Mustafa, Wang Tianye, Xu Mengyue, Zheng Zibo, Al-Qadasi Mohammed, Esmaeeli Omid, Rahim Mohamed, Pakulski Grzegorz, Schmid Jens, Barrios Pedro, Jiang Weihong, Morison Hugh, Mitchell Matthew, Guan Xun, Jaeger Nicolas A F, Rusch Leslie A, Shekhar Sudip, Shi Wei, Yu Siyuan, Cai Xinlun, Chrostowski Lukas
Department of Electrical and Computer Engineering, The University of British Columbia, Vancouver, British Columbia, Canada.
State Key Laboratory of Optoelectronic Materials and Technologies, School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou, Guangdong, China.
Nat Commun. 2024 Oct 21;15(1):9081. doi: 10.1038/s41467-024-53261-x.
Photonics offers a transformative approach to artificial intelligence (AI) and neuromorphic computing by enabling low-latency, high-speed, and energy-efficient computations. However, conventional photonic tensor cores face significant challenges in constructing large-scale photonic neuromorphic networks. Here, we propose a fully integrated photonic tensor core, consisting of only two thin-film lithium niobate (TFLN) modulators, a III-V laser, and a charge-integration photoreceiver. Despite its simple architecture, it is capable of implementing an entire layer of a neural network with a computational speed of 120 GOPS, while also allowing flexible adjustment of the number of inputs (fan-in) and outputs (fan-out). Our tensor core supports rapid in-situ training with a weight update speed of 60 GHz. Furthermore, it successfully classifies (supervised learning) and clusters (unsupervised learning) 112 × 112-pixel images through in-situ training. To enable in-situ training for clustering AI tasks, we offer a solution for performing multiplications between two negative numbers.
光子学通过实现低延迟、高速和节能计算，为人工智能（AI）和神经形态计算提供了一种变革性方法。然而，传统的光子张量核在构建大规模光子神经形态网络方面面临重大挑战。在此，我们提出一种完全集成的光子张量核，它仅由两个薄膜铌酸锂（TFLN）调制器、一个III-V族激光器和一个电荷积分光接收器组成。尽管其架构简单，但它能够以120 GOPS的计算速度实现神经网络的一整层，同时还允许灵活调整输入（扇入）和输出（扇出）的数量。我们的张量核支持以60 GHz的权重更新速度进行快速原位训练。此外，它通过原位训练成功地对112×112像素的图像进行分类（监督学习）和聚类（无监督学习）。为了实现用于聚类AI任务的原位训练，我们提供了一种用于执行两个负数之间乘法的解决方案。
