Ben Recht

Professor at UC Berkeley

Recht helped develop methods for recovering missing information from incomplete datasets and for training machine-learning models efficiently across many processors. His later work tests whether benchmark improvements survive changes in real-world data and examines the limits of automated decision-making.