CGXplain: 基于规则的深度神经网络解释，使用双线性程序（英文版）

英文标题：CGXplain: Rule-Based Deep Neural Network Explanations Using Dual Linear Programs中文摘要：介绍了 CGX（Column Generation eXplainer），一种基于线性规划的分解方法，可以从深度神经网络的隐藏表示中提取规则，优化任意数量的目标，并满足准确复现、稳定性和复杂度降低等特点，提高了可解释性。英文摘要：Rule-based surrogate models are an effective and interpretable way toapproximate a Deep Neural Network's (DNN) decision boundaries, allowing humansto easily understand deep learning models. Current state-of-the-artdecompositional methods, which are those that consider the DNN's latent spaceto