Language Model Adaption for Reinforcement Learning with Natural Language Action Space.
Jiangxing WangJiachen LiXiao HanDeheng YeZongqing LuPublished in: ACL (1) (2024)
Keyphrases
- language model
- action space
- reinforcement learning
- natural language
- state space
- markov decision processes
- language modeling
- state and action spaces
- continuous state
- n gram
- real valued
- probabilistic model
- document retrieval
- action selection
- context sensitive
- continuous state spaces
- information retrieval
- retrieval model
- machine learning
- query expansion
- ad hoc information retrieval
- function approximators
- mixture model
- test collection
- function approximation
- markov decision process
- state action
- query terms
- natural language processing
- stochastic processes
- co occurrence
- multi agent
- optimal policy
- translation model
- learning algorithm
- feature selection
- reward function
- bayesian networks
- single agent
- information extraction
- non stationary