GAM Coach: Towards Interactive and User-centered Algorithmic Recourse

Abstract

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/.

arXiv commentAccepted to CHI 2023. 20 pages, 12 figures. For a demo video, see https://youtu.be/ubacP34H9XE. For a live demo, visit https://poloclub.github.io/gam-coach/

References in corpus (21)

• XGBoost: A Scalable Tree Boosting System
• Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
• Explaining Explanations in AI
• The What-If Tool: Interactive Probing of Machine Learning Models
• Actionable Recourse in Linear Classification
• CNN Explainer: Learning Convolutional Neural Networks with Interactive Visualization
• InterpretML: A Unified Framework for Machine Learning Interpretability
• The Hidden Assumptions Behind Counterfactual Explanations and Principal Reasons
• GAN Lab: Understanding Complex Deep Generative Models using Interactive Visual Experimentation
• Incentivizing High Quality Crowdwork
• Beyond Expertise and Roles: A Framework to Characterize the Stakeholders of Interpretable Machine Learning and their Needs
• Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers
• Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems
• Direct-Manipulation Visualization of Deep Networks
• Algorithmic recourse under imperfect causal knowledge: a probabilistic approach
• Beyond Individualized Recourse: Interpretable and Interactive Summaries of Actionable Recourses
• TimberTrek: Exploring and Curating Sparse Decision Trees with Interactive Visualization
• Interpretability, Then What? Editing Machine Learning Models to Reflect Human Knowledge and Values
• GeCo: Quality Counterfactual Explanations in Real Time
• In Pursuit of Interpretable, Fair and Accurate Machine Learning for Criminal Recidivism Prediction
• Revealing Neural Network Bias to Non-Experts Through Interactive Counterfactual Examples

Cited by in corpus (3)

• MeMemo: On-device Retrieval Augmentation for Private and Personalized Text Generation
• Exploiting Preference Elicitation in Interactive and User-centered Algorithmic Recourse: An Initial Exploration
• AI-Spectra: A Visual Dashboard for Model Multiplicity to Enhance Informed and Transparent Decision-Making