Pith. sign in

REVIEW 1 cited by

Understanding Gradient Clipping in Private SGD: A Geometric Perspective

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 2006.15429 v2 pith:MZ7XEQV2 submitted 2020-06-27 cs.LG cs.CRmath.OCstat.ML

classification cs.LGcs.CRmath.OCstat.ML
keywords gradientclippingprivatebiaslearningconvergencedistributiondistributions
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Deep learning models are increasingly popular in many machine learning applications where the training data may contain sensitive information. To provide formal and rigorous privacy guarantee, many learning systems now incorporate differential privacy by training their models with (differentially) private SGD. A key step in each private SGD update is gradient clipping that shrinks the gradient of an individual example whenever its L2 norm exceeds some threshold. We first demonstrate how gradient clipping can prevent SGD from converging to stationary point. We then provide a theoretical analysis that fully quantifies the clipping bias on convergence with a disparity measure between the gradient distribution and a geometrically symmetric distribution. Our empirical evaluation further suggests that the gradient distributions along the trajectory of private SGD indeed exhibit symmetric structure that favors convergence. Together, our results provide an explanation why private SGD with gradient clipping remains effective in practice despite its potential clipping bias. Finally, we develop a new perturbation-based technique that can provably correct the clipping bias even for instances with highly asymmetric gradient distributions.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Random Batch Method with Momentum Correction

    math.NA 2024-12 reject novelty 5.0 of 10

    A momentum-corrected random batch method is proposed and claimed to reduce the error of the standard random batch method for singular interaction kernels.

Pith tools