🤖 AI Summary
This work proposes a data-aware sensitivity analysis framework for decision tree ensembles, addressing the lack of interpretable evidence grounded in realistic data distributions in existing approaches. By constraining perturbed samples to remain close to the training data distribution, the method generates more realistic and interpretable sensitivity examples to assess model robustness and fairness. It achieves, for the first time, data-aware verification of large-scale, multi-class tree ensembles—such as those with depth 8 and 800 trees—while strengthening the theoretical NP-hard foundation of sensitivity verification. Integrating mixed-integer linear programming (MILP), SMT encoding, and data-aware search optimization, the approach significantly improves verification efficiency and scalability, enabling reliable analysis in high-stakes applications.
📝 Abstract
Decision tree ensembles are widely used in critical domains, making robustness and sensitivity analysis essential to their trustworthiness. We study the feature sensitivity problem, which asks whether an ensemble is sensitive to a specified subset of features -- such as protected attributes -- whose manipulation can alter model predictions. Existing approaches often yield examples of sensitivity that lie far from the training distribution, limiting their interpretability and practical value. We propose a data-aware sensitivity framework that constrains the sensitive examples to remain close to the dataset, thereby producing realistic and interpretable evidence of model weaknesses. To this end, we develop novel techniques for data-aware search using a combination of mixed-integer linear programming (MILP) and satisfiability modulo theories (SMT) encodings. Our contributions are fourfold. First, we strengthen the NP-hardness result for sensitivity verification, showing it holds even for trees of depth 1. Second, we develop MILP-optimizations that significantly speed up sensitivity verification for single ensembles and for the first time can also handle multiclass tree ensembles. Third, we introduce a data-aware framework generating realistic examples close to the training distribution. Finally, we conduct an extensive experimental evaluation on large tree ensembles, demonstrating scalability to ensembles with up to 800 trees of depth 8, achieving substantial improvements over the state of the art. This framework provides a practical foundation for analyzing the reliability and fairness of tree-based models in high-stakes applications.