arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2102.09972cs.LGcs.AIcs.NEstat.ML
收藏

张量分解中的隐正则化

Implicit Regularization in Tensor Factorization

Tel Aviv University(特拉维夫大学)

Noam Razin, Asaf Maman, Nadav Cohen

摘要:

最近在深度学习中解开隐式正则化之谜的努力导致了对矩阵分解的理论关注-通过线性神经网络完成矩阵。作为迈向实际深度学习的一步,我们提供了张量因式分解中隐式正则化的第一个理论分析-通过某种类型的非线性神经网络实现张量完备化。我们规避了臭名昭著的困难张量问题,通过采用动力系统的角度来看,和表征的演变引起的梯度下降。的表征表明,贪婪的低张量秩搜索的形式,我们严格证明在某些条件下,和经验证明在其他情况下。受张量秩捕获非线性神经网络的隐式正则化的启发,我们将其作为复杂性的度量进行了经验探索,并发现它捕获了神经网络概括的数据集的本质。这让我们相信,张量秩可以为解释深度学习中的隐式正则化以及将这种隐式正则化转化为泛化的真实数据的属性铺平道路。

英文摘要:

Recent efforts to unravel the mystery of implicit regularization in deep learning have led to a theoretical focus on matrix factorization -- matrix completion via linear neural network. As a step further towards practical deep learning, we provide the first theoretical analysis of implicit regularization in tensor factorization -- tensor completion via certain type of non-linear neural network. We circumvent the notorious difficulty of tensor problems by adopting a dynamical systems perspective, and characterizing the evolution induced by gradient descent. The characterization suggests a form of greedy low tensor rank search, which we rigorously prove under certain conditions, and empirically demonstrate under others. Motivated by tensor rank capturing the implicit regularization of a non-linear neural network, we empirically explore it as a measure of complexity, and find that it captures the essence of datasets on which neural networks generalize. This leads us to believe that tensor rank may pave way to explaining both implicit regularization in deep learning, and the properties of real-world data translating this implicit regularization to generalization.

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。

↑