Optimal Recourse Summaries via Bi-Objective Decision Tree Learning

2026-05-08

AI Summary

Existing remediation approaches struggle to simultaneously support global auditing and bias detection due to the high cost and inconsistency of aggregating individual-level recommendations, as well as the absence of a systematic model for balancing effectiveness against cost. This work formalizes remediation-aware summarization as a bi-objective optimal decision tree problem, leveraging group partitioning and shared action policies within a shallow axis-aligned decision tree framework to compute the Pareto frontier. This enables flexible post-hoc selection of trade-off points without retraining. The proposed method, SOGAR, consistently outperforms existing approaches in both remediation effectiveness and cost efficiency, while demonstrating high stability, low computational overhead, and strong overall performance.

Abstract

Actionable Recourse provides individuals with actions they can take to change an unfavorable classifier outcome. While useful at the instance level, it is ill-suited for global auditing and bias detection, since aggregating local actions is costly and often inconsistent. Recourse Summaries address this limitation by partitioning the population and assigning one shared action per subgroup, enabling comparison across subgroups. Designing summaries involves a fundamental trade-off between recourse effectiveness and recourse cost, which existing methods do not adequately address. We introduce Summaries of Optimal and Global Actionable Recourse (SOGAR), which formulates recourse summary learning as an optimal decision tree learning problem and finds the Pareto front -- the complete set of solutions where improving one objective necessarily worsens the other. SOGAR enables post-hoc selection of the desired trade-off without retraining. Using shallow axis-parallel decision trees and sparse leaf actions, SOGAR produces stable, low-cost, and effective recourse summaries that outperform existing approaches across effectiveness and cost metrics.

Problem

• actionable recourse
• recourse summaries
• cost-effectiveness trade-off
• global auditing
• bias detection

Innovation

• actionable recourse
• recourse summaries
• bi-objective optimization
• optimal decision trees
• Pareto front