光子贝叶斯神经网络的约束协同设计
Constrained Co-Design for Photonic Bayesian Neural Networks
浏览论文内容
中文总结 AI 辅助
针对经典神经网络的过度自信预测缺陷，研究人员将光子概率计算与贝叶斯神经网络结合，通过约束随机变分推理得出协同设计指南，验证了硬件感知训练可恢复性能及不确定性质量。
中文摘要 AI 辅助
经典神经网络对模糊数据或分布外（OOD）数据常产生过度自信的预测，这种缺陷在安全关键的现实场景中部署AI系统时会愈发明显。贝叶斯神经网络（BNNs）提供了一种基于不确定性感知预测的原则性框架，它用概率分布替代确定性参数，但重复采样会增加延迟、内存流量和能耗。光子概率计算利用固有的光学随机性实现快速并行采样，是一种有前景的替代方案。然而，光子BNNs并非理想的采样器：量化、编程误差、动态范围以及可表示的均值和方差等模拟约束限制了硬件中可实现的变分族。在本研究中，我们探究了哪些硬件施加的约束限制了可扩展的光子BNN推理、这些约束如何表示，以及光子BNN在小型概念验证网络之外可容忍哪些范围。我们将光子BNN推理表述为约束随机变分推理，并对随机性位置、随机性模态、量化、编程误差和均值/方差边界进行系统的 ablation 研究。基于这些结果，我们得出了具体的协同设计指南，区分了可通过训练补偿的硬件约束与需要硬件或架构干预的约束。我们在Dirty-MNIST、CIFAR-10和CINIC-10上，以及以Fashion-MNIST和SVHN作为OOD基准，在耦合的、符合硬件实际的约束下验证了这些指南，结果表明，只要所需的变分族仍可表示，硬件感知训练就能恢复预测性能和不确定性质量，而表示极限的违反则需要针对性的硬件修改。
英文摘要
Classical neural networks frequently produce overconfident predictions on ambiguous or out-of-distribution (OOD) data, a liability that grows with each AI system deployed in safety-critical real-world scenarios. Bayesian neural networks (BNNs) provide a principled framework for uncertainty-aware prediction by replacing deterministic parameters with probability distributions, but repeated sampling increases latency, memory traffic, and energy consumption. Photonic probabilistic computing offers a promising alternative by exploiting intrinsic optical stochasticity for fast and parallel sampling. However, photonic BNNs are not ideal samplers: analog constraints on quantization, programming error, dynamic range, and representable mean and variance restrict the variational families that can be implemented in hardware. In this work, we study which hardware-imposed constraints limit scalable photonic BNN inference, how these constraints can be represented, and which ranges can be tolerated by photonic BNNs beyond small proof-of-concept networks. We formulate photonic BNN inference as constrained stochastic variational inference and perform a systematic ablation study over stochasticity location, stochasticity modality, quantization, programming error, and mean/variance bounds. From these results, we derive concrete co-design guidelines that distinguish hardware constraints that can be compensated by training from those requiring hardware or architecture intervention. We validate these guidelines under coupled, hardware-realistic constraints on Dirty-MNIST, CIFAR-10, and CINIC-10, using Fashion-MNIST and SVHN as OOD benchmarks, showing that hardware-aware training recovers predictive performance and uncertainty quality whenever the required variational family remains representable, whereas violations of representational limits require targeted hardware modifications.
