We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Implicit Bias of Gradient Descent for Mean Squared Error Regression with Two-Layer Wide Neural Networks

Abstract: We investigate gradient descent training of wide neural networks and the corresponding implicit bias in function space. For univariate regression, we show that the solution of training a width-$n$ shallow ReLU network is within $n^{- 1/2}$ of the function which fits the training data and whose difference from the initial function has the smallest 2-norm of the second derivative weighted by a curvature penalty that depends on the probability distribution that is used to initialize the network parameters. We compute the curvature penalty function explicitly for various common initialization procedures. For instance, asymmetric initialization with a uniform distribution yields a constant curvature penalty, and thence the solution function is the natural cubic spline interpolation of the training data. \hj{For stochastic gradient descent we obtain the same implicit bias result.} We obtain a similar result for different activation functions. For multivariate regression we show an analogous result, whereby the second derivative is replaced by the Radon transform of a fractional Laplacian. For initialization schemes that yield a constant penalty function, the solutions are polyharmonic splines. Moreover, we show that the training trajectories are captured by trajectories of smoothing splines with decreasing regularization strength.
Comments: 97 pages, 14 figures. Added the discussion of SGD and implications to generalization
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
MSC classes: 68Q32, 68T05
ACM classes: I.2.6; G.3
Cite as: arXiv:2006.07356 [stat.ML]
  (or arXiv:2006.07356v3 [stat.ML] for this version)

Submission history

From: Hui Jin [view email]
[v1] Fri, 12 Jun 2020 17:46:40 GMT (1686kb,D)
[v2] Sun, 25 Jul 2021 17:16:46 GMT (2818kb,D)
[v3] Wed, 4 Jan 2023 03:10:02 GMT (5307kb,D)

Link back to: arXiv, form interface, contact.