Reproduce “Who Bears the Burden? Learning Responsibility for Shared Constraints in Multi-Agent Reinforcement 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
Xiaoyang Cao, Jingqi Li, Zhe Fu, Alexandre M. Bayen
LiRA reallocates pressure from a shared constraint across agents by learning responsibility shares of a common penalty, without changing rewards or evaluation budgets. Under restrictive assumptions, the paper shows shares can select equilibria with different welfare. In four-agent MetaDrive, LiRA had higher mean welfare and fewer safety events than PAL, but this measured comparison does not guarantee improvement across settings.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.