# Preserving Cross-Modal Stability for Visual Unlearning in Multimodal Scenarios

arXiv: https://arxiv.org/abs/2509.23895 · HTML: https://mlsanity.com/paper/2509.23895/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Machine unlearning, one of 12 directions in LLM safety and security.
- Builds on 18 of its 30 references that are in the map.
- First posted on arXiv 28 Sep 2025.

## Placed in

- LLM safety and security / [Machine unlearning](https://mlsanity.com/safety/machine-unlearning/)

## Builds on (most cited first, 18 of 18)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Membership Inference Attacks against Machine Learning Models](https://mlsanity.com/paper/1610.05820/) (2016)
- [Machine Unlearning](https://mlsanity.com/paper/1912.03817/) (2019)
- [Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networks](https://mlsanity.com/paper/1911.04933/) (2019)
- [Unrolling SGD: Understanding Factors Influencing Machine Unlearning](https://mlsanity.com/paper/2109.13398/) (2021)
- [Towards Unbounded Machine Unlearning](https://mlsanity.com/paper/2302.09880/) (2023)
- [Machine Unlearning of Features and Labels](https://mlsanity.com/paper/2108.11577/) (2021)
- [Can Bad Teaching Induce Forgetting? Unlearning in Deep Networks using an Incompetent Teacher](https://mlsanity.com/paper/2205.08096/) (2022)
- [Fast Machine Unlearning Without Retraining Through Selective Synaptic Dampening](https://mlsanity.com/paper/2308.07707/) (2023)
- [On the Necessity of Auditable Algorithmic Definitions for Machine Unlearning](https://mlsanity.com/paper/2110.11891/) (2021)
- [CentralNet: a Multilayer Approach for Multimodal Fusion](https://mlsanity.com/paper/1808.07275/) (2018)
- [Boosting Multi-modal Model Performance with Adaptive Gradient Modulation](https://mlsanity.com/paper/2308.07686/) (2023)
- [PUMA: Performance Unchanged Model Augmentation for Training Data Removal](https://mlsanity.com/paper/2203.00846/) (2022)
- [Classifier-guided Gradient Modulation for Enhanced Multimodal Learning](https://mlsanity.com/paper/2411.01409/) (2024)
- [MultiDelete for Multimodal Machine Unlearning](https://mlsanity.com/paper/2311.12047/) (2023)
- [Contrastive Unlearning: A Contrastive Approach to Machine Unlearning](https://mlsanity.com/paper/2401.10458/) (2024)
- [On the Separability of Classes with the Cross-Entropy Loss Function](https://mlsanity.com/paper/1909.06930/) (2019)

## Abstract (arXiv)

Visual modality is the most vulnerable to privacy leakage in real-world multimodal applications like autonomous driving with visual and radar data; Machine unlearning removes specific training data from pre-trained models to address privacy leakage, however, existing methods fail to preserve cross-modal knowledge and maintain intra-class structural stability of retain data, leading to reduced overall and other modalities' performance during visual unlearning; to address these challenges, we propose a Cross-modal Contrastive Unlearning (CCU) framework, which integrates three key components: (a) selective visual unlearning: employing inverse contrastive learning to dissociate visual representations from their original semantics, (b) cross-modal knowledge retention: preserving other modalities' discriminability through semantic consistency, and (c) dual-set contrastive separation: preserving the model performance via isolation of structural perturbations between the unlearn set and retain set; extensive experiments on three datasets demonstrate the superiority of CCU, and our method achieves a 7.12% accuracy improvement with only 7% of the unlearning time compared to the top-accuracy baseline.