Ma Shi-Yuan, Wang Tianyu, Laydevant Jérémie, Wright Logan G, McMahon Peter L
School of Applied and Engineering Physics, Cornell University, Ithaca, NY, USA.
USRA Research Institute for Advanced Computer Science, Mountain View, CA, USA.
Nat Commun. 2025 Jan 3;16(1):359. doi: 10.1038/s41467-024-55220-y.
Energy efficiency in computation is ultimately limited by noise, with quantum limits setting the fundamental noise floor. Analog physical neural networks hold promise for improved energy efficiency compared to digital electronic neural networks. However, they are typically operated in a relatively high-power regime so that the signal-to-noise ratio (SNR) is large (>10), and the noise can be treated as a perturbation. We study optical neural networks where all layers except the last are operated in the limit that each neuron can be activated by just a single photon, and as a result the noise on neuron activations is no longer merely perturbative. We show that by using a physics-based probabilistic model of the neuron activations in training, it is possible to perform accurate machine-learning inference in spite of the extremely high shot noise (SNR ~ 1). We experimentally demonstrated MNIST handwritten-digit classification with a test accuracy of 98% using an optical neural network with a hidden layer operating in the single-photon regime; the optical energy used to perform the classification corresponds to just 0.038 photons per multiply-accumulate (MAC) operation. Our physics-aware stochastic training approach might also prove useful with non-optical ultra-low-power hardware.
计算中的能量效率最终受噪声限制，量子极限设定了基本噪声底限。与数字电子神经网络相比，模拟物理神经网络有望提高能量效率。然而，它们通常在相对高功率状态下运行，以便信噪比（SNR）较大（>10），且噪声可被视为微扰。我们研究光学神经网络，其中除最后一层外的所有层都在每个神经元仅能被单个光子激活的极限下运行，结果神经元激活时的噪声不再仅仅是微扰。我们表明，通过在训练中使用基于物理的神经元激活概率模型，尽管存在极高的散粒噪声（SNR ~ 1），仍有可能进行准确的机器学习推理。我们通过实验展示了使用具有在单光子状态下运行的隐藏层的光学神经网络进行MNIST手写数字分类，测试准确率达98%；用于执行分类的光能量对应于每次乘加（MAC）操作仅0.038个光子。我们的物理感知随机训练方法可能对非光学超低功耗硬件也有用。
