Interpretability credit debate points back to Doshi-Velez & Kim's foundational 2017 paper

StephenLCasper · x · 2026-09-07

On X, Stephen Casper was credited for a sequence on interpretability research, but he redirected credit to earlier work—specifically Finale Doshi-Velez and Been Kim's 2017 position paper Towards A Rigorous Science of Interpretable Machine Learning (arXiv:1702.08608).

The paper defines interpretability, discusses when it is (and isn't) needed, and proposes a taxonomy for rigorous evaluation, noting the field still lacks consensus on what interpretability is and how to measure it.

Original post →

More from Research

Research channel →