Proxy tasks and subjective measures can be misleading in evaluating explainable AI systems.

Zana Buinca, Phoebe Lin, Krzysztof Z. Gajos, Elena L. Glassman

Browse the full IUI paper archive.