ICED performs interpretable concept-level unlearning in VLMs by constructing a concept vocabulary via MLLM and decomposing visual representations for targeted optimization.
Machine unlearning via task simplex arithmetic,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
ICED: Concept-level Machine Unlearning via Interpretable Concept Decomposition
ICED performs interpretable concept-level unlearning in VLMs by constructing a concept vocabulary via MLLM and decomposing visual representations for targeted optimization.