arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2004.13954cs.LGcs.AIstat.ML
收藏

重新思考DNN的泛化、泛化和谱偏差之间的关系

Rethink the Connections among Generalization, Memorization and the Spectral Bias of DNNs

Huazhong University of Science and Technology(华中科技大学);Baidu Research(百度研究院)

Xiao Zhang, Haoyi Xiong, Dongrui Wu

摘要:

具有足够记忆随机噪声能力的过参数化深度神经网络(DNN)可以实现出色的泛化性能,挑战经典学习理论中的偏差-方差权衡。最近的研究表明,DNN首先学习简单的模式,然后记住噪声;其他一些工作表明,DNN在训练过程中从低频到高频学习目标函数时存在频谱偏差。然而,我们发现学习偏差的单调性并不总是成立:在深度双下降的实验设置下,DNN的高频分量在训练的后期减少,导致测试误差的第二次下降。此外,我们发现DNN的谱可以用来指示测试误差的第二次下降,即使它是仅从训练集计算的。

英文摘要:

Over-parameterized deep neural networks (DNNs) with sufficient capacity to memorize random noise can achieve excellent generalization performance, challenging the bias-variance trade-off in classical learning theory. Recent studies claimed that DNNs first learn simple patterns and then memorize noise; some other works showed a phenomenon that DNNs have a spectral bias to learn target functions from low to high frequencies during training. However, we show that the monotonicity of the learning bias does not always hold: under the experimental setup of deep double descent, the high-frequency components of DNNs diminish in the late stage of training, leading to the second descent of the test error. Besides, we find that the spectrum of DNNs can be applied to indicating the second descent of the test error, even though it is calculated from the training set only.

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。

↑