English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Do Influence-Derived Data Perturbations Enable Machine Unlearning? A Reproduction Audit of Deep Perturbation Learning

Forum topic · 小凯 · 2026-09-15

Summary

This paper evaluates Deep Perturbation Learning (DPL), a method that perturbs training images and labels along influence-derived directions, in three roles previously claimed for machine unlearning: direct deletion signal, utility-preserving regularizer, and warm start for adversarial unlearning. The authors audit the public implementation and find two correctness issues: image perturbation directions are computed on augmented, normalized tensors but applied to raw images, and label perturbations fall below float32 resolution, leaving labels effectively unchanged. After correcting the image-perturbation pipeline, DPL fails the direct deletion criterion on CIFAR-10/ResNet-18 across three paired seeds with exact-seed retraining baselines. Its utility impact is inconsistent across seeds, and once direction-computation time is included it underperforms a simple warm-start baseline. Single-seed Tiny ImageNet checks likewise do not support DPL as a regularizer or warm start. The results cover random instance deletion only and do not rule out influence methods in other unlearning regimes. The authors release a role-matched evaluation protocol and an audit checklist for perturbation-based unlearning claims (arXiv:2609.12313).

This post on zhichai.net summarizes a machine learning paper (arXiv: 2609.12313) by Chenkai Wu, Chrispine Kambimbi, Qinyang Zeng, and Jun Yan, published 2026-09-15.

Key points

  • What is evaluated: Deep Perturbation Learning (DPL), which perturbs training images and labels along influence-derived directions, in three roles prior work positioned it for machine unlearning:
  • 1. A direct deletion signal (the strongest claim) 2. A utility-preserving regularizer 3. A warm start for adversarial unlearning
  • Motivation: Evidence for the weaker roles had been used to support the stronger one, so the authors test each role separately under a matched protocol with exact-seed retraining baselines.
  • Implementation audit findings (two correctness issues in the public code):
  • Image directions are computed on augmented, normalized tensors but applied to raw images.
  • Label perturbation magnitude falls below float32 resolution, so labels are effectively unchanged.
  • Main results after correcting the image-perturbation pipeline:
  • DPL fails the direct deletion criterion on CIFAR-10/ResNet-18 across three paired seeds.
  • Its utility impact is inconsistent in sign across seeds.
  • Once direction-computation time is counted, DPL underperforms a simple warm-start baseline.
  • Single-seed Tiny ImageNet checks do not support DPL as a regularizer or warm start either; preprocessing inconsistencies in the released code prevent definitive direct comparison.
  • Scope caveat: Results cover random instance deletion only and do not rule out the effectiveness of influence-based methods in other unlearning regimes.
  • Artifacts: The authors release a role-matched evaluation protocol and an audit checklist for perturbation-based unlearning claims.

Tags

#machine-unlearning#deep-perturbation-learning#reproducibility#influence-functions#audit#cifar-10#deep-learning

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634844