The What-If Tool: Interactive Probing of Machine Learning Models

TLDR

A key challenge in developing and deploying machine learning systems is understanding their performance across a wide range of inputs. To address this challenge, we created the What‑If Tool, an open‑source application that allows practitioners to probe, visualize, and analyze ML systems with minimal coding. The What‑If Tool lets practitioners test performance in hypothetical scenarios, assess feature importance, compare multiple models and data subsets, measure fairness metrics, and provides a design description and real‑life usage reports. Real‑life usage of the What‑If Tool was reported at several organizations.

Abstract

A key challenge in developing and deploying Machine Learning (ML) systems is understanding their performance across a wide range of inputs. To address this challenge, we created the What-If Tool, an open-source application that allows practitioners to probe, visualize, and analyze ML systems, with minimal coding. The What-If Tool lets practitioners test performance in hypothetical situations, analyze the importance of different data features, and visualize model behavior across multiple models and subsets of input data. It also lets practitioners measure systems according to multiple ML fairness metrics. We describe the design of the tool, and report on real-life usage at different organizations.