通过解耦的直通估计器改进离散优化
Improving Discrete Optimisation Via Decoupled Straight-Through Estimator
AI总结：
本文提出了解耦直通估计器，通过分别控制前向传递和反向传递的温度参数，提高了离散优化的性能。
摘要：
直通估计器（STE）是训练包含离散变量的神经网络的主要方法，通过使用可微的替代品将梯度传递通过以实现基于梯度的优化。然而，现有的STE变体将两种根本不同的关注点混为一谈：前向传递的随机性，它控制探索和潜在空间利用，以及反向传递的梯度分散，即学习信号如何分布在各个类别之间。我们证明这些关注点在本质上是不同的，并且将它们绑定到单一温度参数会错失显著的性能提升。我们提出了解耦的直通（解耦ST），这是一种最小的修改，它引入了前向传递（τ_f）和反向传递（τ_b）分别的温度参数。这种简单的改变使得可以独立调整探索和梯度分散。在三个不同的任务（随机二进制网络、分类自动编码器和可微逻辑门网络）中，解耦ST始终优于身份STE、Softmax STE和直通Gumbel-Softmax。关键的是，最优（τ_f，τ_b）配置远离对角线τ_f = τ_b，证实了这两个关注点确实需要不同的答案，并且单温度方法在本质上受到限制。
英文摘要：
The Straight-Through Estimator (STE) is the dominant method for training neural networks with discrete variables, enabling gradient-based optimisation by routing gradients through a differentiable surrogate. However, existing STE variants conflate two fundamentally distinct concerns: forward-pass stochasticity, which controls exploration and latent space utilisation, and backward-pass gradient dispersion i.e how learning signals are distributed across categories. We show that these concerns are qualitatively different and that tying them to a single temperature parameter leaves significant performance gains untapped. We propose Decoupled Straight-Through (Decoupled ST), a minimal modification that introduces separate temperatures for the forward pass ($τ_f$) and the backward pass ($τ_b$). This simple change enables independent tuning of exploration and gradient dispersion. Across three diverse tasks (Stochastic Binary Networks, Categorical Autoencoders, and Differentiable Logic Gate Networks), Decoupled ST consistently outperforms Identity STE, Softmax STE, and Straight-Through Gumbel-Softmax. Crucially, optimal $(τ_f, τ_b)$ configurations lie far off the diagonal $τ_f = τ_b$, confirming that the two concerns do require different answers and that single-temperature methods are fundamentally constrained.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
