arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2111.12876stat.MLcs.LGmath.OCmath.PR
收藏

非凸环境下SGLD的时间无关泛化界

Time-independent Generalization Bounds for SGLD in Non-convex Settings

University of Oxford(牛津大学)

Tyler Farghly, Patrick Rebeschini

摘要:

我们建立了泛化误差界的随机梯度朗之万动力学(SGLD)的耗散性和光滑性的假设下,具有恒定的学习率,已受到越来越多的关注,在采样/优化文献的设置。与非凸设置中SGLD的现有界限不同,我们的是时间无关的,并且随着样本量的增加而衰减到零。使用一致稳定的框架,我们建立时间无关的界限,利用Wasserstein收缩性质的朗之万扩散,这也使我们能够规避需要绑定梯度使用Lipschitz假设。我们的分析还支持SGLD的变体,这些变体使用不同的离散化方法,结合欧几里得投影或使用非各向同性噪声。

英文摘要:

We establish generalization error bounds for stochastic gradient Langevin dynamics (SGLD) with constant learning rate under the assumptions of dissipativity and smoothness, a setting that has received increased attention in the sampling/optimization literature. Unlike existing bounds for SGLD in non-convex settings, ours are time-independent and decay to zero as the sample size increases. Using the framework of uniform stability, we establish time-independent bounds by exploiting the Wasserstein contraction property of the Langevin diffusion, which also allows us to circumvent the need to bound gradients using Lipschitz-like assumptions. Our analysis also supports variants of SGLD that use different discretization methods, incorporate Euclidean projections, or use non-isotropic noise.

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。

↑