Reproduce “GatedVLA with Camera-Based Emergency Braking: AlpaSim E2E Closed Loop Challenge Technical Report” 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
Kaneyoshi Hiratsuka, Naoki Takada, Makoto Kawano
GatedVLA combines a visual attention mask with a separate camera-based emergency-braking rule for closed-loop driving. The mask guides the action expert toward selected image evidence, while braking intervenes when the policy fails to slow for a lead obstacle. Across 150 local scenes, mean score rose from 0.517 with the mask to 0.599 with braking; thresholds were tuned on those scenes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.