A Unifying View on Implicit Bias in Training Linear Neural Networks

We study the implicit bias of gradient flow (i.e., gradient descent with infinitesimal step size) on linear neural network training. We propose a tensor formulation of neural networks that includes fully-connected, diagonal, and convolutional networks as special cases, and investigate the linear ver...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2021-09
Main Authors: Yun, Chulhee, Krishnan, Shankar, Mobahi, Hossein
Format: Article
Language:English
Subjects:
Bias
Gradient flow
Mathematical analysis
Neural networks
Norms
Regression analysis
Tensors
Training
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We study the implicit bias of gradient flow (i.e., gradient descent with infinitesimal step size) on linear neural network training. We propose a tensor formulation of neural networks that includes fully-connected, diagonal, and convolutional networks as special cases, and investigate the linear version of the formulation called linear tensor networks. With this formulation, we can characterize the convergence direction of the network parameters as singular vectors of a tensor defined by the network. For \(L\)-layer linear tensor networks that are orthogonally decomposable, we show that gradient flow on separable classification finds a stationary point of the \(\ell_{2/L}\) max-margin problem in a "transformed" input space defined by the network. For underdetermined regression, we prove that gradient flow finds a global minimum which minimizes a norm-like function that interpolates between weighted \(\ell_1\) and \(\ell_2\) norms in the transformed input space. Our theorems subsume existing results in the literature while removing standard convergence assumptions. We also provide experiments that corroborate our analysis.
ISSN:2331-8422