通过潜在聚类发现扩散基于的反事实概念方向
Discovering Concept Directions from Diffusion-based Counterfactuals via Latent Clustering
AI总结：
本文提出CDLC方法，通过潜在聚类提取概念方向，减少存储需求并加速概念发现，适用于高风险领域和多样化数据模态。
摘要：
基于概念的解释已经 emerged 作为可解释人工智能中的有效方法，通过将模型决策与人类可理解的概念对齐，从而获得可解释的见解。然而，现有方法依赖于计算密集型的流程，并且难以高效地捕捉复杂的、语义性的概念。本文介绍了通过潜在聚类（CDLC）发现概念方向，该方法通过聚类从真实和扩散生成的反事实图像对中提取的潜在差异向量，提取出全局的、类特定的概念方向。CDLC相比基线方法减少了约4.6%的存储需求，并加速了概念发现约5.3%，而无需GPU进行聚类，从而在潜在维度上实现了高效提取多维语义概念。该方法在真实世界皮肤病变数据集上得到验证，证明所提取的概念方向与临床认可的皮肤病学特征一致，并在某些情况下揭示了数据集特定的偏见或未知生物标志物。这些结果表明，CDLC是可解释的、可扩展的，并且适用于高风险领域和多样的数据模态。
英文摘要：
Concept-based explanations have emerged as an effective approach within Explainable Artificial Intelligence, enabling interpretable insights by aligning model decisions with human-understandable concepts. However, existing methods rely on computationally intensive procedures and struggle to efficiently capture complex, semantic concepts. This work introduces the Concept Directions via Latent Clustering (CDLC), which extracts global, class-specific concept directions by clustering latent difference vectors derived from factual and diffusion-generated counterfactual image pairs. CDLC reduces storage requirements by ~4.6% and accelerates concept discovery by ~5.3% compared to the baseline method, while requiring no GPU for clustering, thereby enabling efficient extraction of multidimensional semantic concepts across latent dimensions. This approach is validated on a real-world skin lesion dataset, demonstrating that the extracted concept directions align with clinically recognized dermoscopic features and, in some cases, reveal dataset-specific biases or unknown biomarkers. These results highlight that CDLC is interpretable, scalable, and applicable across high-stakes domains and diverse data modalities.
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。
