Drift Q-Learning overview

Drift Q-Learning

Preprint 2026

DriftQL generates actions in one forward pass using a learned drift field. Attraction keeps candidates near the data, repulsion spreads them out, and the critic tilts the objective toward higher value. No ODE solver, no distillation network.

2 min · Anas Houssaini, Mohamad H. Danesh, Amin Abyaneh, Scott Fujimoto, Hsiu-Chin Lin, David Meger