CHOQOLATE：使用 Choquet 积分组织概念瓶颈潜在空间
CHOQOLATE: Organizing Concept Bottleneck Latent Spaces with Choquet Integrals
AI总结：
提出基于2-可加Choquet积分的可解释层CHOQOLATE，合并相关概念以改善CBM的忠实性，在四个数据集上实现准确率与可解释性的良好权衡，并支持无需组标注的测试时干预。
AI中文摘要：
基于 CLIP 等视觉-语言模型构建的概念瓶颈模型（CBM）将潜在空间表示为人类可理解的概念。这些表示并不忠实：相关概念相互纠缠，因此单个分数无法反映其预期含义。我们提出 CHOQOLATE，一种基于 2-可加 Choquet 积分的可解释性设计层，它将相关概念合并为紧凑节点。在四个数据集上，CHOQOLATE 实现了良好的准确率-可解释性权衡，具有权重稀疏且语义连贯的节点。由实验支持的闭式梯度推导解释了为何 Choquet 层能在无需显式监督的情况下驱动这种组织。Choquet 权重还可直接映射到 Shapley 值，从而实现测试时干预。在标准偏差缓解基准上，训练后抑制虚假概念的表现与需要组标注或重新训练的方法相当，且两者都不需要。
英文摘要：
Concept Bottleneck Models (CBMs) built on vision-language models such as CLIP represent a latent space as human-understandable concepts. These representations are unfaithful: related concepts are entangled, so individual scores do not reflect their intended meaning. We propose CHOQOLATE, an interpretable-by-design layer based on 2-additive Choquet integrals, which merges correlated concepts into compact nodes. Across four datasets, CHOQOLATE achieves a favorable accuracy-interpretability trade-off, with weight-sparse and semantically coherent nodes. A closed-form gradient derivation, backed by experiments, explains why Choquet layers drive this organization without explicit supervision. Choquet weights also map directly to Shapley values, which enables test-time intervention. On standard bias-mitigation benchmarks, suppressing spurious concepts after training performs on par with methods that require group annotations or retraining, while needing neither.
