arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.11752cs.LGcs.AImath.OC
收藏

对角化SGD:通过重参数化与平滑实现非可微模型的快速且收敛的SGD

Diagonalisation SGD: Fast & Convergent SGD for Non-Differentiable Models via Reparameterisation and Smoothing

NTU Singapore(新加坡南洋理工大学);Jane Street(简街资本)

Dominik Wagner, Basim Khajwal, C. -H. Luke Ong

AI总结:

本文提出对角化SGD,通过分段定义非可微函数并逐步提高平滑近似精度,在保证收敛到原始目标稳定点的同时实现低方差、快速稳定的优化。

摘要:

众所周知,重参数化梯度估计器在实践中表现出低方差,但对于非可微模型是有偏的。这可能损害基于梯度的优化方法(如随机梯度下降SGD)的正确性。我们引入一个简单的语法框架来分段定义非可微函数,并提出一种系统化方法来获得使重参数化梯度估计器无偏的平滑化。我们的主要贡献是SGD的一种新变体,即对角化随机梯度下降(Diagonalisation Stochastic Gradient Descent),它在优化过程中逐步提高平滑近似的精度,并且我们证明了收敛到未平滑(原始)目标的稳定点。我们的实证评估揭示了相对于现有最优方法的优势:我们的方法简单、快速、稳定,并在工作归一化方差上实现了数量级的降低。

英文摘要:

It is well-known that the reparameterisation gradient estimator, which exhibits low variance in practice, is biased for non-differentiable models. This may compromise correctness of gradient-based optimisation methods such as stochastic gradient descent (SGD). We introduce a simple syntactic framework to define non-differentiable functions piecewisely and present a systematic approach to obtain smoothings for which the reparameterisation gradient estimator is unbiased. Our main contribution is a novel variant of SGD, Diagonalisation Stochastic Gradient Descent, which progressively enhances the accuracy of the smoothed approximation during optimisation, and we prove convergence to stationary points of the unsmoothed (original) objective. Our empirical evaluation reveals benefits over the state of the art: our approach is simple, fast, stable and attains orders of magnitude reduction in work-normalised variance.

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。

↑