Reproduce “ResWAM: Learning What Changes, Not What Stays, for World-Action Models” 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 Fan, Gaowen Liu, Ramana Rao Kompella, Jason Corso, Sijia Liu

ResWAM jointly predicts robot actions and changes in future visual features, using the difference between current and future features from a frozen vision encoder. This change-focused target has higher feature rank and more spatially focused responses. On RoboTwin 2.0’s 50-task test, it reaches 93.8% success versus 90.9% for complete-future prediction at identical compute; evidence is benchmark-specific.

Summary by alphaXiv.

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