Reproduce “Context-Aware Sparse Deep Coordination Graphs” 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
Tonghan Wang, Liang Zeng, Weijun Dong, Qianlan Yang, Yang Yu, Chongjie Zhang
CASEC learns sparse coordination graphs for cooperative multi-agent reinforcement learning by ranking pairwise links according to how much their payoff varies across another agent’s possible actions. It retains a fraction of highest-ranked links for Max-Sum action selection, while learning action representations. On Sensor, the selected threshold saved 90% of messages; thresholds were fixed, and sparse graphs can yield suboptimal actions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.