Reproduce “When Does Cross-View Consistency Matter in Multi-View Appearance Augmentation? A Controlled Study for Visuomotor Policies” 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
Zixuan Jiang
The authors test how pairing visual changes across two camera views affects robot manipulation. They replay fixed robot states, preserving each camera’s perturbation distribution, trajectories, and action labels while changing cross-view pairing. For a late-fusion Diffusion Policy on seven LIBERO tasks, anti-coupled geometry lowered success by 3.30 percentage points; tested appearance changes had small costs, not proof of no harm.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.