arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2303.00883cs.LGmath.OCstat.ML
收藏

用于非凸优化的方差缩减裁剪

Variance-reduced Clipping for Non-convex Optimization

MIT–IBM Watson AI Lab(麻省理工学院-IBM沃森人工智能实验室);IBM Research(IBM研究院);Massachusetts Institute of Technology(麻省理工学院)

Amirhossein Reisizadeh, Haochuan Li, Subhro Das, Ali Jadbabaie

AI总结:

本文针对$(L_0,L_1)$平滑非凸优化,将SPIDER方差缩减与裁剪学习率结合,把随机梯度复杂度改进至阶最优的$O(ε^{-3})$及有限和情形的$O(\sqrt{n}ε^{-2}+n)$,并在视觉任务上验证性能。

摘要:

梯度裁剪是深度学习应用(例如大规模语言建模)中用于缓解梯度爆炸的标准训练技术。近期实验研究表明,在使用梯度裁剪进行训练时,训练目标沿其轨迹的平滑性表现出一种相当特殊的行为。也就是说,平滑性随梯度范数增长。这与民间非凸优化中广泛确立的假设(即$L$——平滑性)形成鲜明对比,在该假设中,平滑性被假定在全局范围内由常数$L$界定。最近提出的$(L_0,L_1)$——平滑性是一种更宽松的概念,可捕捉非凸优化中的此类行为。具体而言,已有研究表明,在这种放宽的平滑性假设下,带裁剪的SGD需要$O(ε^{-4})$次随机梯度计算才能找到一个$ε$——平稳解。本文采用一种名为SPIDER的方差缩减技术,并证明对于精心设计的学习率,该复杂度可提升至$O(ε^{-3})$,且达到阶最优。我们设计的学习率包含裁剪技术,以缓解不断增长的平滑性。此外,当目标函数是$n$个分量的平均值时,我们将现有的随机梯度复杂度界$O(nε^{-2})$改进为$O(\sqrt{n} ε^{-2} + n)$,这同样达到阶最优。除理论最优性外,采用我们所设计参数的SPIDER在若干视觉任务中相较SVRG和SARAH等方差缩减方法展现出相当的实证性能。

英文摘要:

Gradient clipping is a standard training technique used in deep learning applications such as large-scale language modeling to mitigate exploding gradients. Recent experimental studies have demonstrated a fairly special behavior in the smoothness of the training objective along its trajectory when trained with gradient clipping. That is, the smoothness grows with the gradient norm. This is in clear contrast to the well-established assumption in folklore non-convex optimization, a.k.a. $L$--smoothness, where the smoothness is assumed to be bounded by a constant $L$ globally. The recently introduced $(L_0,L_1)$--smoothness is a more relaxed notion that captures such behavior in non-convex optimization. In particular, it has been shown that under this relaxed smoothness assumption, SGD with clipping requires $O(ε^{-4})$ stochastic gradient computations to find an $ε$--stationary solution. In this paper, we employ a variance reduction technique, namely SPIDER, and demonstrate that for a carefully designed learning rate, this complexity is improved to $O(ε^{-3})$ which is order-optimal. Our designed learning rate comprises the clipping technique to mitigate the growing smoothness. Moreover, when the objective function is the average of $n$ components, we improve the existing $O(nε^{-2})$ bound on the stochastic gradient complexity to $O(\sqrt{n} ε^{-2} + n)$, which is order-optimal as well. In addition to being theoretically optimal, SPIDER with our designed parameters demonstrates comparable empirical performance against variance-reduced methods such as SVRG and SARAH in several vision tasks.

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。

↑