纳秒
光子学
人工神经网络
计算机科学
上下文图像分类
人工智能
光电子学
材料科学
图像(数学)
模式识别(心理学)
物理
激光器
光学
作者
Amirreza Shoobi,Alexander J. Geers,Anish Mondal,Kaisarbek Omirzakhov,Farshid Ashtiani,Firooz Aflatouni
出处
期刊:Custom Integrated Circuits Conference
[Institute of Electrical and Electronics Engineers]
日期:2025-04-13
卷期号:: 1-6
被引量:5
标识
DOI:10.1109/cicc63670.2025.10983432
摘要
Photonic deep neural networks can perform fast and energy-efficient classification through computation by propagation and low-loss interconnects. Here, a review of our work on the first integrated end-to-end photonic-electronic deep neural network for direct image classification that achieved a classification time of 570 ps is presented. Then, the first monolithically integrated, reconfigurable, and trainable optical nonlinear activation function (NLAF) designed in a 45nm CMOS SOI process for scalable silicon photonic deep neural networks is introduced. The NLAF supports reconfigurable ReLU and Sigmoid functionalities over a 6 GHz of bandwidth, with trainable parameters for turn-on thresholds, slopes, and saturation levels to effectively mitigate process, voltage, and temperature (PVT) variations, ensuring robustness in deep networks by reducing the number of trained parameters by orders of magnitude. The chip was implemented using a 45nm CMOS SOI process within a footprint of 0.028 mm2and achieves a minimum activation power of -23 dBm.
科研通智能强力驱动
Strongly Powered by AbleSci AI