让相关概念在模型中彻底分离，提升可解释性与可控性
Post-Hoc Concept Disentanglement: From Correlated to Isolated Concept Representations
- 通过非正交损失强制学习正交概念方向
- 实现概念插入与移除时互不干扰，效果优于传统CAV
- 适合需要精确控制模型语义的可解释性研究
概念激活向量（CAVs）广泛用于将人类可理解的概念建模为神经网络潜在空间中的方向。该方法通过对比概念样本与非概念样本的激活来识别方向，但当概念存在相关性（如CelebA数据集中“胡须”与“领带”常共现）时，会生成相似且非正交的方向，导致概念难以独立解释，并在激活控制任务中产生不良影响。为此，本文提出一种后处理概念解耦方法，引入非正交性损失，在保持方向正确性的前提下，促进概念方向的正交化。我们在CelebA和合成的FunnyBirds数据集上，使用VGG16和ResNet18进行评估。结果表明，正交化后的概念表示在激活控制任务中表现更优：（1）可借助生成模型将孤立概念插入输入图像；（2）能有效移除目标概念，同时对相关概念影响更小，优于基线CAV方法。
原文摘要 · Abstract (English)
Concept Activation Vectors (CAVs) are widely used to model human-understandable concepts as directions within the latent space of neural networks. They are trained by identifying directions from the activations of concept samples to those of non-concept samples. However, this method often produces similar, non-orthogonal directions for correlated concepts, such as "beard" and "necktie" within the CelebA dataset, which frequently co-occur in images of men. This entanglement complicates the interpretation of concepts in isolation and can lead to undesired effects in CAV applications, such as activation steering. To address this issue, we introduce a post-hoc concept disentanglement method that employs a non-orthogonality loss, facilitating the identification of orthogonal concept directions while preserving directional correctness. We evaluate our approach with real-world and controlled correlated concepts in CelebA and a synthetic FunnyBirds dataset with VGG16 and ResNet18 architectures. We further demonstrate the superiority of orthogonalized concept representations in activation steering tasks, allowing (1) the insertion of isolated concepts into input images through generative models and (2) the removal of concepts for effective shortcut suppression with reduced impact on correlated concepts in comparison to baseline CAVs.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。