Implicit Regularization in Tensor Factorization

Noam Razin · Asaf Maman · Nadav Cohen

Keywords: [ Deep Learning Theory ]

[ Abstract ]
[ Slides [ Paper ] [ Visit Poster at Spot A4 in Virtual World ]
Thu 22 Jul 9 a.m. PDT — 11 a.m. PDT
Spotlight presentation: Deep Learning Theory 6
Thu 22 Jul 5 a.m. PDT — 6 a.m. PDT


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.

Chat is not available.