Single-Agent vs. Multi-Agent Techniques for Concurrent Reinforcement Learning of Negotiation Dialogue Policies.
Kallirroi GeorgilaClaire NelsonDavid R. TraumPublished in: ACL (1) (2014)
Keyphrases
- single agent
- multi agent
- reinforcement learning
- hierarchical reinforcement learning
- multi agent coordination
- optimal policy
- interaction protocols
- action space
- control policies
- policy iteration algorithm
- multi agent systems
- multiagent reinforcement learning
- multiple agents
- reinforcement learning agents
- partially observable markov decision processes
- cooperative
- learning agents
- multiagent systems
- multi agent environments
- reward function
- dec pomdps
- control policy
- model free
- function approximation
- markov decision process
- window search
- continuous state
- markov decision processes
- decision problems
- autonomous agents
- software agents
- reinforcement learning algorithms
- stochastic games
- state abstraction
- state space
- markov decision problems
- learning agent
- learning algorithm
- finite state
- machine learning
- multiagent learning
- reinforcement learning methods
- temporal difference
- partially observable
- dynamic programming