GAM Coach: Towards Interactive and User-centered Algorithmic Recourse
This work addresses the need for more user-centered and interactive algorithmic recourse in high-stakes domains, though it is incremental as it builds on existing recourse techniques with a focus on customization and visualization.
The authors tackled the problem of generating actionable recourse plans for machine learning predictions by developing GAM Coach, a system that uses integer linear programming and interactive visualizations to create customizable counterfactual explanations for Generalized Additive Models, and a user study with 41 participants showed users preferred personalized plans and found the tool usable and useful.
Machine learning (ML) recourse techniques are increasingly used in high-stakes domains, providing end users with actions to alter ML predictions, but they assume ML developers understand what input variables can be changed. However, a recourse plan's actionability is subjective and unlikely to match developers' expectations completely. We present GAM Coach, a novel open-source system that adapts integer linear programming to generate customizable counterfactual explanations for Generalized Additive Models (GAMs), and leverages interactive visualizations to enable end users to iteratively generate recourse plans meeting their needs. A quantitative user study with 41 participants shows our tool is usable and useful, and users prefer personalized recourse plans over generic plans. Through a log analysis, we explore how users discover satisfactory recourse plans, and provide empirical evidence that transparency can lead to more opportunities for everyday users to discover counterintuitive patterns in ML models. GAM Coach is available at: https://poloclub.github.io/gam-coach/.