arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2011.06702cs.LGmath.OC
收藏

神经网络训练技术正则化优化轨迹:一项实证研究

Neural Network Training Techniques Regularize Optimization Trajectory: An Empirical Study

University of Utah(犹他大学);The Ohio State University(俄亥俄州立大学)

Cheng Chen, Junjie Yang, Yi Zhou

AI总结:

本文通过实证与理论分析表明,非线性激活、批归一化和跳跃连接等训练技术可依据规律性原则约束模型更新方向,使其对齐优化轨迹并提升 DNN 非凸优化的收敛性。

摘要:

现代深度神经网络(DNN)训练会使用多种训练技术,例如非线性激活函数、batch normalization(批归一化)、skip-connections(跳跃连接)等。尽管这些技术十分有效,但它们在实践中究竟如何帮助加速 DNN 训练,仍然是一个未解之谜。本文对这些训练技术在 DNN 优化中的正则化效应进行了实证研究。具体而言,我们发现成功的 DNN 训练所产生的优化轨迹始终遵循某种规律性原则,该原则对模型更新方向进行正则化,使其与轨迹方向保持一致。理论上,我们证明这种规律性原则能够为非凸优化带来收敛保证,并且收敛速度取决于一个正则化参数。实证上,我们发现应用了这些训练技术的 DNN 训练能够快速收敛,并以较大的正则化参数遵循该规律性原则,这意味着模型更新与轨迹具有良好的一致性。另一方面,未使用这些训练技术的 DNN 训练收敛缓慢,并且只以较小的正则化参数遵循该规律性原则,这意味着模型更新与轨迹未能良好对齐。因此,不同训练技术通过该规律性原则对模型更新方向进行正则化,从而促进收敛。

英文摘要:

Modern deep neural network (DNN) trainings utilize various training techniques, e.g., nonlinear activation functions, batch normalization, skip-connections, etc. Despite their effectiveness, it is still mysterious how they help accelerate DNN trainings in practice. In this paper, we provide an empirical study of the regularization effect of these training techniques on DNN optimization. Specifically, we find that the optimization trajectories of successful DNN trainings consistently obey a certain regularity principle that regularizes the model update direction to be aligned with the trajectory direction. Theoretically, we show that such a regularity principle leads to a convergence guarantee in nonconvex optimization and the convergence rate depends on a regularization parameter. Empirically, we find that DNN trainings that apply the training techniques achieve a fast convergence and obey the regularity principle with a large regularization parameter, implying that the model updates are well aligned with the trajectory. On the other hand, DNN trainings without the training techniques have slow convergence and obey the regularity principle with a small regularization parameter, implying that the model updates are not well aligned with the trajectory. Therefore, different training techniques regularize the model update direction via the regularity principle to facilitate the convergence.

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。

↑