Pith. sign in

REVIEW 6 cited by

On the Origin of Implicit Regularization in Stochastic Gradient Descent

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2101.12176 v1 pith:FW2YBU67 submitted 2021-01-28 cs.LG stat.ML

On the Origin of Implicit Regularization in Stochastic Gradient Descent

classification cs.LG stat.ML
keywords learninglossrategradientimplicitbatchsmalltest
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

For infinitesimal learning rates, stochastic gradient descent (SGD) follows the path of gradient flow on the full batch loss function. However moderately large learning rates can achieve higher test accuracies, and this generalization benefit is not explained by convergence bounds, since the learning rate which maximizes test accuracy is often larger than the learning rate which minimizes training loss. To interpret this phenomenon we prove that for SGD with random shuffling, the mean SGD iterate also stays close to the path of gradient flow if the learning rate is small and finite, but on a modified loss. This modified loss is composed of the original loss function and an implicit regularizer, which penalizes the norms of the minibatch gradients. Under mild assumptions, when the batch size is small the scale of the implicit regularization term is proportional to the ratio of the learning rate to the batch size. We verify empirically that explicitly including the implicit regularizer in the loss can enhance the test accuracy when the learning rate is small.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Convergence of difference inclusions via a diameter criterion

    math.OC 2026-05 unverdicted novelty 7.0

    A diameter criterion tied to a potential function certifies convergence of difference inclusions, enabling discrete proofs for first-order optimization methods with diminishing steps.

  2. Estimating Implicit Regularization in Deep Learning

    stat.ML 2026-05 unverdicted novelty 7.0

    Gradient matching empirically recovers implicit regularization effects such as l2 penalties from early stopping and dropout in neural networks.

  3. Avoiding unsafe sets when training with Langevin Dynamics

    cs.LG 2026-07 accept novelty 6.0

    Langevin training trajectories on strongly convex losses avoid geometrically isolated failure regions with probability exponentially small in dimension after an O(d) burn-in, with a local spectral rate controlling tra...

  4. Second-Order Path Kernel Interpolation Formulas in Machine Learning

    cs.LG 2026-06 unverdicted novelty 6.0

    Derives second-order path-kernel interpolation formulas for gradient descent, SGD, and momentum training, adding curvature terms and a concentration estimate around the expected prediction.

  5. Thermodynamic Irreversibility of Training Algorithms

    cond-mat.stat-mech 2026-05 unverdicted novelty 6.0

    Four characterizations of irreversibility in training algorithms are equivalent to leading order in step size and produce an emergent force that breaks reparametrization symmetries while favoring minimum entropy produ...

  6. Gradient-Descent Steps to Success over Mean Accuracy: A Paradigm Shift for ML

    cs.LG 2026-06 unverdicted novelty 5.0

    The paper defines computational effort as the number of gradient descent steps to reach target accuracy with high probability, shows large learning rates minimize this effort across models, and identifies phase transi...