HydroGym lands on Nature cover: 60+ RL environments to unify fluid flow control research

jiqizhixin · x · 2026-09-20

Researchers from the University of Washington, University of Michigan, and collaborators published HydroGym on the cover of Nature — a solver-agnostic reinforcement learning platform addressing the lack of shared benchmarks in fluid dynamics control.

The team calls it the field's "ImageNet moment" for active flow control.

Original post →

More from Research

Research channel →