用信息论找出点云中影响模型判断的关键可解释结构。
InfoCons: Identifying Interpretable Critical Concepts in Point Clouds via Information Theory
- 基于信息论分解点云，识别出对预测有因果影响的3D概念。
- 在合成数据上优于四个基线方法，能准确捕捉关键点集。
- 适用于自动驾驶等安全场景，帮助诊断模型失败原因。
随着点云模型在自动驾驶等安全关键场景中的应用日益广泛，其可解释性变得至关重要。本文聚焦于将模型输出归因于可解释的关键概念，即输入点云中有意义的子集。理想的關鍵子集应具备两个特性：忠实性（保留对预测具有因果影响的点）和概念一致性（形成与人类感知一致的语义结构）。我们提出InfoCons，一种基于信息论原理的解释框架，通过引入可学习先验，将点云分解为3D概念，并评估其对模型预测的因果影响。我们在合成分类数据集上对InfoCons进行了定性和定量评估，对比了四种基线方法。进一步实验表明，该方法在两个真实世界数据集上具有良好的可扩展性和灵活性，并成功应用于两个利用点云关键评分的任务中。
原文摘要 · Abstract (English)
Interpretability of point cloud (PC) models becomes imperative given their deployment in safety-critical scenarios such as autonomous vehicles. We focus on attributing PC model outputs to interpretable critical concepts, defined as meaningful subsets of the input point cloud. To enable human-understandable diagnostics of model failures, an ideal critical subset should be *faithful* (preserving points that causally influence predictions) and *conceptually coherent* (forming semantically meaningful structures that align with human perception). We propose InfoCons, an explanation framework that applies information-theoretic principles to decompose the point cloud into 3D concepts, enabling the examination of their causal effect on model predictions with learnable priors. We evaluate InfoCons on synthetic datasets for classification, comparing it qualitatively and quantitatively with four baselines. We further demonstrate its scalability and flexibility on two real-world datasets and in two applications that utilize critical scores of PC.
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。