Gradient Concentration, Not Weight Saliency, Explains Representation-Level Class Unlearning

• URL: http://arxiv.org/abs/2607.21353v1
• Date: Thu, 23 Jul 2026 14:17:51 GMT
• Title: Gradient Concentration, Not Weight Saliency, Explains Representation-Level Class Unlearning
• Abstract summary: We perform the first controlled ablation of the saliency masking mechanism used by SalUn.<n>We find that representation-level forgetting is governed by gradient concentration and representation geometry rather than by the specific identity of saliency-selected weights.
• Score: 3.33554314544636
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Machine unlearning aims to remove the influence of specific training data while preserving model utility. Many state-of-the-art approaches pursue this goal by restricting the forgetting update to a subset of parameters selected through gradient-based saliency. Although such methods are widely adopted, the actual contribution of saliency-based weight selection to representation-level forgetting remains unclear. In this work, we perform the first controlled ablation of the saliency masking mechanism used by SalUn. Using a matched-compute experimental design on CIFAR-10 and CIFAR-100 with ResNet-18, we compare saliency-based masking against random masks of equal sparsity and unconstrained updates, while keeping the unlearning objective, optimization schedule, and computational budget fixed. Across multiple representation-level evaluations, including linear probing, prototype recovery, and layer-wise CKA, the three configurations exhibit statistically equivalent representation-level recoverability. We find that forget gradients are strongly concentrated in the final network layers (approximately 92% of the squared gradient energy on CIFAR-10) before any mask is applied, causing all masking strategies to operate within the same representational subspace. Furthermore, saliency masks show limited class specificity (specificity index 0.09-0.11), selecting highly overlapping parameter subsets across different forget classes. Our findings suggest that, in the studied setting, representation-level forgetting is primarily governed by gradient concentration and representation geometry rather than by the specific identity of saliency-selected weights. More broadly, the results support a growing body of evidence indicating that effective representation-level unlearning requires objectives that act directly on latent representations rather than on increasingly sophisticated weight-selection strategies.

Related papers

• One Layer's Trash is Another Layer's Treasure: Adaptive Layer-wise Visual Token Selection in LVLMs [18.48496973561215]
  Large Vision-Language Models (LVLMs) have achieved remarkable success across diverse multimodal tasks, yet their practical deployment remains constrained by the computational burden arising from lengthy visual tokens.<n>We propose Adaptive Layer-wise Visual Token Selection (ALVTS), a novel framework that breaks away from the conventional static token pruning paradigm.<n>With an 89% token compression ratio, ALVTS retains 96.7% of the original model's accuracy, achieving a superior efficiency-accuracy trade-off for LVLM inference.
  arXiv  Detail & Related papers  (2026-06-12T08:58:58Z)
• Self-Distillation of Hidden Layers for Self-Supervised Representation Learning [18.042078225633045]
  We introduce Bootleg, a method that bridges the divide by tasking the model with predicting latent representations from hidden layers of a teacher network.<n>We demonstrate that Bootleg significantly outperforms comparable baselines on classification of ImageNet-1K and iNaturalist-21, and semantic segmentation of ADE20K and Cityscapes.
  arXiv  Detail & Related papers  (2026-03-16T17:13:27Z)
• POUR: A Provably Optimal Method for Unlearning Representations via Neural Collapse [12.913395960667161]
  In computer vision, machine unlearning aims to remove the influence of specific visual concepts or training images without retraining from scratch.<n>We extend the notion of unlearning to the representation level, deriving a three-term interplay between forgetting efficacy, retention fidelity, and class separation.<n>Experiments on CIFAR-10/100 and PathMNIST demonstrate that POUR achieves effective unlearning while preserving retained knowledge, outperforming state-of-the-art unlearning methods on both classification-level and representation-level metrics.
  arXiv  Detail & Related papers  (2025-11-24T17:38:53Z)
• Seeing What Matters: Empowering CLIP with Patch Generation-to-Selection [54.21851618853518]
  We present a concise yet effective approach called Patch Generation-to-Selection to enhance CLIP's training efficiency.<n>Our approach, CLIP-PGS, sets new state-of-the-art results in zero-shot classification and retrieval tasks.
  arXiv  Detail & Related papers  (2025-03-21T12:10:38Z)
• SeWA: Selective Weight Average via Probabilistic Masking [51.015724517293236]
  We show that only a few points are needed to achieve better and faster convergence.<n>We transform the discrete selection problem into a continuous subset optimization framework.<n>We derive the SeWA's stability bounds, which are sharper than that under both convex image checkpoints.
  arXiv  Detail & Related papers  (2025-02-14T12:35:21Z)
• Bridge the Points: Graph-based Few-shot Segment Anything Semantically [79.1519244940518]
  Recent advancements in pre-training techniques have enhanced the capabilities of vision foundation models. Recent studies extend the SAM to Few-shot Semantic segmentation (FSS) We propose a simple yet effective approach based on graph analysis.
  arXiv  Detail & Related papers  (2024-10-09T15:02:28Z)
• Language Models as Zero-shot Lossless Gradient Compressors: Towards General Neural Parameter Prior Models [56.00251589760559]
  Large language models (LLMs) can act as gradient priors in a zero-shot setting.<n>We introduce LM-GC, a novel method that integrates LLMs with arithmetic coding.<n>Experiments indicate that LM-GC surpasses existing state-of-the-art lossless compression methods.
  arXiv  Detail & Related papers  (2024-09-26T13:38:33Z)
• Class Gradient Projection For Continual Learning [99.105266615448]
  Catastrophic forgetting is one of the most critical challenges in Continual Learning (CL) We propose Class Gradient Projection (CGP), which calculates the gradient subspace from individual classes rather than tasks.
  arXiv  Detail & Related papers  (2023-11-25T02:45:56Z)
• TSG: Target-Selective Gradient Backprop for Probing CNN Visual Saliency [72.9106103283475]
  We study the visual saliency, a.k.a. visual explanation, to interpret convolutional neural networks. Inspired by those observations, we propose a novel visual saliency framework, termed Target-Selective Gradient (TSG) backprop. The proposed TSG consists of two components, namely, TSG-Conv and TSG-FC, which rectify the gradients for convolutional layers and fully-connected layers, respectively.
  arXiv  Detail & Related papers  (2021-10-11T12:00:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.