More Centralized Training, Still Decentralized Execution: Multi-Agent Conditional Policy Factorization.
Jiangxing WangDeheng YeZongqing LuPublished in: CoRR (2022)
Keyphrases
- multi agent
- cooperative
- peer to peer
- multi agent systems
- optimal policy
- single agent
- action selection
- training process
- matrix factorization
- intelligent agents
- agent oriented
- distributed environment
- test set
- online learning
- distributed systems
- training set
- supervised learning
- probabilistic model
- pairwise
- digital libraries
- training phase
- reinforcement learning
- machine learning
- neural network