Visual Analysis of Discrimination in Machine Learning

The growing use of automated decision-making in critical applications, such as crime prediction and college admission, has raised questions about fairness machine learning. How can we decide whether different treatments are reasonable or discriminatory? In this paper, investigate discrimination learning from a visual analytics perspective propose an interactive visualization tool, DiscriLens, to support more comprehensive analysis. To reveal detailed information on algorithmic discrimination, DiscriLens identifies collection potentially discriminatory itemsets based causal modeling classification rules mining. By combining extended Euler diagram with matrix-based visualization, develop novel set facilitate the exploration interpretation itemsets. A user study shows that users interpret visually encoded quickly accurately. Use cases demonstrate provides informative guidance understanding reducing discrimination.