ICED: Concept-level Machine Unlearning via Interpretable Concept Decomposition

TL;DR

This paper introduces a concept-level unlearning method for vision-language models that enables precise removal of specific knowledge while preserving unrelated semantics, using interpretable concept decomposition.

Contribution

It proposes a novel framework that constructs a task-specific concept vocabulary and decomposes visual representations for fine-grained unlearning in VLMs.

Findings

Enables more comprehensive target forgetting.

Better preserves non-target knowledge within images.

Maintains competitive model utility.

Abstract

Machine unlearning in Vision-Language Models (VLMs) is typically performed at the image or instance level, making it difficult to precisely remove target knowledge without affecting unrelated semantics. This issue is especially pronounced since a single image often contains multiple entangled concepts, including both target concepts to be forgotten and contextual information that should be preserved. In this paper, we propose an interpretable concept-level unlearning framework for VLMs, which constructs a compact task-specific concept vocabulary from the forgetting set using a multimodal large language model. In addition to modality alignment, visual representations are decomposed into sparse, nonnegative combinations of semantic concepts, providing an explicit interface for fine-grained knowledge manipulation. Based on this decomposition, our method formulates unlearning as concept-level optimization, where target concepts are selectively suppressed while intra-instance non-target semantics and global cross-modal knowledge are preserved. Extensive experiments across both in-domain and out-of-domain forgetting settings demonstrate that our method enables more comprehensive target forgetting, better preserves non-target knowledge within the same image, and maintains competitive model utility compared with existing VLM unlearning methods.