MANCE: 流形感知的概念擦除

摘要

概念擦除旨在从表示中移除目标概念，同时保留其中编码的其他信息。这具有一定难度，因为表示编码了许多通常与擦除目标相关的概念，因此移除目标可能对这些概念造成损害。我们提出流形约束假说（MCH）：若自然表示集中于结构化的低维流形，则干预应受限于该流形，从而在干预过程中更好地保留表示中编码的其他信息。我们将MCH实例化为一种新的概念擦除方法：流形感知概念擦除（MANCE）。MANCE利用预测目标概念的分类器提供的信号，对表示进行迭代更新。我们通过自然输入获得的表示估计流形，然后将概念移除的更新投影到估计的流形上。我们在涵盖文本和视觉的119个设置中进行了广泛评估，包括13个语言模型、三个NLP概念以及40个CelebA-CLIP属性。在先前方法之上应用MANCE始终取得了更好的泄露改善结果。我们还引入了MANCE+和MANCE++，它们在应用MANCE之前先使用封闭形式的擦除算法，相较于匹配的全空间更新，实现了更好的泄露-精准性权衡。我们的最佳方法MANCE++在非线性概念擦除上达到了最先进的结果。这些结果支持了擦除场景下的MCH：干预应受限于自然表示流形。

English

Concept erasure aims to remove a target concept from a representation while preserving the other information encoded in it. This is difficult because representations encode many concepts that are often correlated with the erasure target, so removing the target risks damaging them. We propose the Manifold Constraint Hypothesis (MCH): if natural representations concentrate on a structured, lower-dimensional manifold, then interventions should be constrained to that manifold and better preserve other information encoded in the representation during interventions. We instantiate MCH in a new concept erasure method: MANifold aware Concept Erasure (MANCE). MANCE performs iterative updates to the representations using signals from a classifier that predicts a target concept. We estimate the manifold using representations obtained from natural inputs, and then we project the concept removal update to the estimated manifold. We perform extensive evaluation on 119 settings spanning text and vision, including 13 language models, three NLP concepts, and 40 CelebA-CLIP attributes. Employing MANCE on top of previous methods shows consistent improved leakage results. We also introduce MANCE+ and MANCE++, which prepend a closed-form erasure algorithm before employing MANCE, achieving better leakage--surgicality tradeoffs relative to matched full-space updates. MANCE++, our best method, achieves state-of-the-art results on nonlinear concept erasure. These results support MCH in the erasure setting: interventions should be constrained to the natural representation manifold.