用k-means重构CNN解释力，让模型自己讲清决策依据。
Post-hoc Self-explanation of CNNs
- 用k-means替代最后线性层，实现后验自解释
- 浅层特征激活生成语义更准的解释图
- 无需梯度、适合想理解模型决策的人
尽管标准卷积神经网络（CNN）可数学重释为自解释模型（SEMs），但其内置原型无法准确代表数据。将最后的线性层替换为基于k-means的分类器可解决此问题，且不损害性能。本文提出统一形式化方法，用于对分类器、编码器最终输出（B4）以及中间特征激活组合进行k-means后验解释。后一种方法利用卷积感受野的空间一致性，生成基于概念的解释图，并由无梯度特征归因图支持。在ResNet34上的实证评估表明，使用较浅层、压缩较少的特征激活（如最后三个块B234）可在语义保真度与预测性能间实现权衡。
原文摘要 · Abstract (English)
Although standard Convolutional Neural Networks (CNNs) can be mathematically reinterpreted as Self-Explainable Models (SEMs), their built-in prototypes do not on their own accurately represent the data. Replacing the final linear layer with a $k$-means-based classifier addresses this limitation without compromising performance. This work introduces a common formalization of $k$-means-based post-hoc explanations for the classifier, the encoder's final output (B4), and combinations of intermediate feature activations. The latter approach leverages the spatial consistency of convolutional receptive fields to generate concept-based explanation maps, which are supported by gradient-free feature attribution maps. Empirical evaluation with a ResNet34 shows that using shallower, less compressed feature activations, such as those from the last three blocks (B234), results in a trade-off between semantic fidelity and a slight reduction in predictive performance.
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。