Reproduce “HumanoidToolBench: Benchmarking Humanoid Tool Use from Selection to Mobile Execution” 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
Kyochul Jang, Seohyeon Park, Ohchul Kwon, Sangjun Park, Junhyeok Choi, Seungyeop Yi, Chaeyun Kim, Sangkyu Lee and 4 more
The authors introduce HumanoidToolBench, a benchmark that tests humanoid robots’ tool selection, pickup, stationary use, and mobile use across three tasks, with and without decoy tools. Results show that tool contact does not ensure task completion: FastWAM’s standard BallMove success falls from 76% at L1 to 9% at L2, where layout also changes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.