Reproduce “Learning Martial Arts Motion Skills for Animating Opponents in VR Sparring Games” 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
Jiqing Wen, Qixuan Shi, Robert LiKamWa, Pavan Turaga
The authors present a Tang Soo Do-inspired pipeline that trains a simulated humanoid on captured martial-arts motion to strike changing targets and evade simulated attacks. Adversarial style rewards and skill discovery shape a reusable repertoire, while task rewards guide behavior. Striking reaches about 98% success at 1.6 m, but this does not establish technical correctness or gameplay quality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.