Reproduce “Active Learning of Inverse Models with Intrinsically Motivated Goal Exploration in Robots” 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
Adrien Baranes, Pierre-Yves Oudeyer
SAGG-RIAC learns inverse models by selecting goals in task space and using goal-directed motor exploration to learn movements that achieve them. Its higher level samples regions by local competence progress, while a lower level refines policies. In a simulated 15-joint arm, it outperformed tested methods at 200,000 micro-actions (p = 0.002). Results depend on a specified, useful low-dimensional task space.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.