Reproduce “Generalists Act, Specialists Intervene: Modular Stage-Selective Reinforcement Learning for Vision-Language-Action Manipulation” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Chongyu Zhu, Hyegang Kim, Jaden Hinds, Juan Sebastian Rojas, Ramy Elmallah, Chi-Guhn Lee

RouteRLT keeps a frozen VLA as default, routing precision-critical phases to learned reinforcement-learning specialists. A GRU selector chooses controllers from compact-state history; handoff logic clears queued actions when control changes. On one physical connector task, the system completed 23 of 35 trials (65.7%), versus 3 of 35 (8.6%) for SmolVLA alone; gains were not uniform across tested backbones.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.