Reproduce “GUILD: Guided Iterative Data Collection for Imitation Learning” 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
Shivin Dass, Kevin Nguyen, Rohan Yelandur, Junhong Xu, Roberto Martín-Martín
GUILD guides robot imitation-learning data collection by proposing plausible tasks and using estimated task utility to steer which demonstrations are collected next. A vision-language model expands candidate tasks, while trajectory attribution estimates their value against a small target-task validation set. In real-world tests, GUILD’s average performance was about 25% higher than VLM-Uniform, though utility is an imperfect proxy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.