More Centralized Training, Still Decentralized Execution: Multi-Agent Conditional Policy Factorization.

Jiangxing Wang, Deheng Ye, Zongqing Lu

Browse the full ICLR paper archive.