Countering Variations and Thermal Effects for Accurate Optical Neural Networks
Countering Variations and Thermal Effects for Accurate Optical Neural Networks
复制标题
DOI：
10.1145/3400302.3415682
复制
发表时间：
2020-11
期刊：
影响因子：
--
通讯作者：
Ying Zhu;Grace Li Zhang;Bing Li;Xunzhao Yin;Cheng Zhuo;Huaxi Gu;Tsung-Yi Ho;Ulf Schlichtmann
中科院分区：
文献类型：
--
作者：
Ying Zhu;Grace Li Zhang;Bing Li;Xunzhao Yin;Cheng Zhuo;Huaxi Gu;Tsung-Yi Ho;Ulf Schlichtmann
光神经网络（ONN）已成为一种有前景的高性能计算平台，可加速深度神经网络。在ONN中，光的相位通过马赫曾德干涉仪（MZI）进行调制，MZI以网格状布局连接以实现乘法累加运算。然而，ONN 对工艺变化和热效应非常敏感。这种敏感性导致 ONN 的推理精度显着下降，从而使其在实践中无法使用。在本文中，我们提出了一个框架来校准过程变化并通过功率补偿来抵消热效应。实验结果表明，所提出的框架可以在变化和热效应下恢复推理精度，例如，Cifar10上LeNet-5的推理精度从低至11.05%恢复到74.11%，使得ONN可以达到与软件训练后的精度相似的推理精度，同时在神经拟态计算中提供高带宽。
Optical neural networks (ONNs) have emerged as a promising high-performance computing platform to accelerate deep neural networks. In ONNs, phases of light are modulated through Mach-Zehnder Interferometers (MZIs), and MZIs are connected in a gridlike layout to implement multiply-accumulate operations. However, ONNs are very sensitive to process variations and thermal effects. This sensitivity leads to a significant degradation of inference accuracy of ONNs and thus renders them unusable in practice. In this paper, we propose a framework to calibrate process variations and counter thermal effects by power compensation. Experimental results demonstrate that the proposed framework can recover the inference accuracy under variations and thermal effects, e.g., from as low as 11.05% back to 74.11% for LeNet-5 on Cifar10, so that ONNs can achieve an inference accuracy similar to the accuracy after software training while providing their high bandwidth in neuromorphic computing.