Pith. sign in

REVIEW 3 cited by

Normalized/Clipped SGD with Perturbation for Differentially Private Non-Convex Optimization

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 2206.13033 v1 pith:YPQ2E7RJ submitted 2022-06-27 cs.LG cs.ITmath.ITstat.ML

Normalized/Clipped SGD with Perturbation for Differentially Private Non-Convex Optimization

classification cs.LG cs.ITmath.ITstat.ML
keywords algorithmsdp-nsgdachieveassumptionsconvergencedeltadp-sgdepsilon
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

By ensuring differential privacy in the learning algorithms, one can rigorously mitigate the risk of large models memorizing sensitive training data. In this paper, we study two algorithms for this purpose, i.e., DP-SGD and DP-NSGD, which first clip or normalize \textit{per-sample} gradients to bound the sensitivity and then add noise to obfuscate the exact information. We analyze the convergence behavior of these two algorithms in the non-convex optimization setting with two common assumptions and achieve a rate $\mathcal{O}\left(\sqrt[4]{\frac{d\log(1/\delta)}{N^2\epsilon^2}}\right)$ of the gradient norm for a $d$-dimensional model, $N$ samples and $(\epsilon,\delta)$-DP, which improves over previous bounds under much weaker assumptions. Specifically, we introduce a regularizing factor in DP-NSGD and show that it is crucial in the convergence proof and subtly controls the bias and noise trade-off. Our proof deliberately handles the per-sample gradient clipping and normalization that are specified for the private setting. Empirically, we demonstrate that these two algorithms achieve similar best accuracy while DP-NSGD is comparatively easier to tune than DP-SGD and hence may help further save the privacy budget when accounting the tuning effort.

discussion (0)

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

Forward citations

Cited by 3 Pith papers

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

  1. Mitigating Membership Inference in Intermediate Representations with Differentially Private Training

    cs.LG 2026-02 unverdicted novelty 6.0

    LM-DP-SGD estimates layer-specific MIA risks from shadow models and reweights gradients to give stronger protection to vulnerable layers, improving the privacy-utility trade-off over uniform DP-SGD.

  2. Revisiting Privacy Amplification by Subsampling in Selective Release DPSGD

    cs.LG 2026-06 unverdicted novelty 5.0

    DPSR-CG corrects the privacy accounting for selective release in DPSGD by addressing sampling probability variation and reports strong empirical results on MNIST, CIFAR-10, IMDB, and FMNIST while claiming strict privacy.

  3. From Gradient Clipping to Structural Refinement: Improving DPSGD for Medical Image Segmentation

    cs.CV 2026-06 unverdicted novelty 4.0

    Evaluating DPSGD clipping methods on medical segmentation shows prior assumptions fail in this domain, but adding morphological refinement and an adaptive DP-Morph variant improves utility under privacy constraints.