VLMs真的能忘记吗？基于免训练的视觉概念反向学习基准测试
Suppression Is Not Forgetting: Residual Recoverability in Visual Concept Unlearning for VLMs
AI总结：
本文提出VLM-UnBench，首个评估VLM免训练视觉概念反向学习的基准，通过四层遗忘等级、七源数据集和十一概念轴，验证提示对概念抑制的真实效果，发现对象和场景概念最难抑制，提示级抑制与真实视觉概念擦除存在明显差距。
摘要：
VLMs在大规模网络数据上训练后保留了敏感和受版权保护的视觉概念，部署时可能需要移除。基于训练的反向学习方法存在结构缺陷：在狭窄的遗忘集上微调会降级一般能力，使无法将后续性能下降归因于反向学习过程本身。基于免训练的方法通过提示或系统指令抑制概念，但缺乏严格评估。本文引入VLM-UnBench，首个评估VLM免训练视觉概念反向学习的基准。它涵盖四层遗忘等级、七源数据集和十一概念轴，并配以三级探测分类法和五种评估条件，以区分真正的遗忘与指令合规。在8种评估设置和13种VLM配置中，现实的反向学习提示使遗忘准确率接近无指令基线；有意义的减少仅在揭示目标概念的oracle条件下出现。对象和场景概念是最难抑制的，即使有明确的遗忘指令，更强的指令调优模型仍保持能力。这些结果揭示了提示级抑制与真实视觉概念擦除之间的明显差距。
英文摘要：
Vision-language models (VLMs) may need to forget visual concepts after deployment because of privacy, copyright, licensing, safety, or policy changes. Conventional machine unlearning modifies model parameters, which may be costly or inaccessible for API-only models. Prompt-based suppression offers a training-free alternative, but does it make a concept inaccessible or merely change the model's answer? We investigate this question in off-the-shelf VLMs. Our visually grounded, multi-probe evaluation first verifies that a model recognizes each concept from the image, then tests its recoverability through multiple-choice, short-answer, and indirect queries. Across objects, scenes, and identities, prompt suppression reduces short-answer recall for some concepts while leaving multiple-choice and indirect performance largely unchanged. Explicitly listing the concepts to suppress often increases short-answer recall, suggesting that the list itself cues the answer. Beyond prompting, decoding constraints, representation editing, and parameter updates can suppress particular responses while the same concept remains detectable through other queries. A model may stop naming a visual concept yet still identify or use it when queried differently.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
