AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Beyond Sample-Level Forgetting: Improving Reliability in Multimodal Unlearning

Jianzhou Wang, Yirui Wu, Lixin Yuan, Wenxiao Zhang, Jun Liu

ICML 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Unified any-modality language models

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a causal-perspective approach to multimodal unlearning that decouples unimodal and multimodal knowledge to improve effectiveness, reliability, and locality.

Problem

Multimodal unlearning, removing specific data from pretrained multimodal models, struggles to achieve effectiveness, reliability, and locality because unimodal and multimodal knowledge are interdependent.

Approach

Introduces a causal perspective and performs multimodal unlearning with decoupled unimodal and multimodal knowledge, moving beyond sample-level forgetting.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add negative pairs (N-Pairs) (ablation)contrastive semantic editing · retrieval privacy 50 samplesMeff: 95.82 -> 99.54Improved
add negative pairs (N-Pairs) (ablation)contrastive semantic editing · retrieval privacy 50 samplesMrel: 96.3 -> 99.33Improved
add positive pairs (P-Pairs) (ablation)contrastive semantic editing · retrieval privacy 50 samplesMeff / Mrel: Meff 86.16 -> 99.54; Mrel 83.79 -> 99.33Improved
add gradient-based optimization (ablation)contrastive semantic editing · retrieval privacy 50 samplesMloc: 66.38 -> 86.09Improved
generate Rj with both U'i and K'i instead of only U'i (ablation)generation strategy · retrieval privacy 50 samplesMrel: 76.60 -> 99.33Improved
generate Rj with both U'i and K'i instead of only K'i (ablation)generation strategy · retrieval privacy 50 samplesMloc: 64.22 -> 86.09Improved
swap in proper U'i (vs [U'j, K'i]) (ablation)generation strategy · retrieval privacy 50 samplesMloc: 62.57 -> 86.09Improved
use structured generated Rj instead of random selecting Xj (ablation)generation strategy · retrieval privacy 50 samplesMeff / Mrel: Meff 98.72 -> 99.54; Mrel 85.35 -> 99.33Improved
random selecting Xj as Rj instead of generated (ablation)generation strategy · retrieval privacy 50 samplesMloc: 89.7 vs 86.09Mixed
increase number of generated images NR1 (ablation)virtual samples · retrieval privacy 50 samplesreliability and localityHurt
increase number of generated texts NR2 (ablation)virtual samples · retrieval privacy 50 samplesperformanceNo clear effect
use adaptive token-wise alpha instead of fixed alpha (ablation)contrastive semantic editing · retrieval privacy 50 samplesMeff / Mrel / MlocImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy