Decision Predicate Graph:为树集成预测给出可验证的结构化局部解释
Where Tree Ensembles Decide: Structural Local Explanation with Decision Predicate Graphs
做表格数据建模、用过 XGBoost 或随机森林又嫌 SHAP 只给权重不给规则的话,这篇很对口。它直接告诉你哪个阈值跨越最能翻转预测,定位植入故障阈值比随机强好几倍。可惜无训练编辑还不能真正纠错,当诊断工具看比较合适。
arXiv 论文提出局部 Decision Predicate Graph(local DPG),把树集成模型在单条样本上执行的全部决策规则聚合成一份可验证的计算记录,并从中选出 critical predicate——即其阈值被跨越后最能推动模型输出转向目标类别的已执行谓词,用户可直接得到具体阈值和共享该阈值的树,无需外挂解释器。
实验覆盖 16 个数据集和四类集成模型族:单次跨越 critical predicate 能在 46-69% 的错误预测中恢复正确类别,而按 TreeSHAP 排序第一的特征使用同一规则仅达 29-50%;在植入故障的测试中,该方法在 46-71% 的错误中定位出故障阈值,随机选择仅 6-21%。它对预测不同的邻近样本的区分能力也优于随机谓词和最近阈值谓词。
论文坦承局限:无训练的谓词编辑在未见数据上并未降低误差。作者定位其为特征归因和反事实解释的补充——归因告诉你哪个特征重要,critical predicate 告诉你预测具体在哪个阈值处可以被改变。
Where Tree Ensembles Decide: Structural Local Explanation with Decision Predicate Graphs
Tree ensembles are among the most accurate models for tabular data, yet the computation behind a single prediction is spread over hundreds of decision rules. Feature attributions and counterfactuals explain such predictions through the influence of features or a change of the input, leaving open which of the thresholds applied by the trees a prediction responds to, while structural explanations compress these rules and lose the trees behind them. To address this gap, we propose a local Decision Predicate Graph (local DPG), which gathers the rules executed for a sample into a verifiable record of the computation, with a class support that reproduces every explained prediction. From these rules we select the critical predicate, the executed predicate whose crossing is estimated to move the model most strongly toward a chosen class, which offers the user a concrete threshold and the trees that share it, without an external explainer. Across 16 datasets and four ensemble families, it separates differently predicted neighbours better than random and nearest-threshold predicates. A single crossing recovers the true class in 46-69% of wrong predictions, against 29-50% when the same rule is applied to the feature ranked first by TreeSHAP, a measure close to the selection criterion, and on planted faults it finds the faulty threshold in 46-71% of the resulting errors, against 6-21% at random. Although training-free edits of these predicates did not reduce the error on unseen data, the critical predicate complements attributions and counterfactuals by naming where a prediction can be changed.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include TreeSHAP, interpretability, counterfactual explanations, tabular data.
Source attribution: curated from arXiv cs.LG (@Sylvio Barbon Junior, Luca Longo).