用两阶段训练法优化光电子神经网络，大幅提速并提效。
Direct Kernel Optimization: Efficient Design for Opto-Electronic Convolutional Neural Networks
- 先训电子CNN，再合成光学卷积核复现其特征
- 同等算力下精度提升一倍，训练时间显著减少
- 适合需高效设计光电子系统的科研与工程人员
混合光电子神经网络结合光学前端与电子后端完成视觉任务，但光学与电子组件的联合端到端（E2E）优化因参数空间大且需反复模拟光学卷积而计算成本高昂。本文提出直接核优化（DKO），一种两阶段训练框架：先训练传统电子卷积神经网络（CNN），再合成光学核以复现第一层卷积滤波器，从而降低优化维度，并在优化过程中避免大量模拟光学卷积。我们在单目深度估计模型上通过仿真评估了DKO，结果表明，在相同计算预算下，其精度为E2E训练的两倍，同时显著缩短训练时间。鉴于混合光电子系统优化的巨大计算挑战，本研究使DKO成为可扩展的训练与实现方法。
原文摘要 · Abstract (English)
Hybrid opto-electronic neural networks combine optical front-ends with electronic back-ends to perform vision tasks, but joint end-to-end (E2E) optimization of optical and electronic components is computationally expensive due to large parameter spaces and repeated optical convolutions. We propose Direct Kernel Optimization (DKO), a two-stage training framework that first trains a conventional electronic CNN and then synthesizes optical kernels to replicate the first-layer convolutional filters, reducing optimization dimensionality and avoiding hefty simulated optical convolutions during optimization. We evaluate DKO in simulation on a monocular depth estimation model and show that it achieves twice the accuracy of E2E training under equal computational budgets while reducing training time. Given the substantial computational challenges of optimizing hybrid opto-electronic systems, our results position DKO as a scalable optimization approach to train and realize these systems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。