Is Reinforcement Learning (Not) for Natural Language Processing?: Benchmarks, Baselines, and Building Blocks for Natural Language Policy Optimization.
Rajkumar RamamurthyPrithviraj AmmanabroluKianté BrantleyJack HesselRafet SifaChristian BauckhageHannaneh HajishirziYejin ChoiPublished in: CoRR (2022)
Keyphrases
- building blocks
- natural language processing
- natural language
- reinforcement learning
- optimal policy
- machine learning
- information extraction
- semantic analysis
- language processing
- knowledge representation
- partially observable environments
- policy search
- text processing
- optimization problems
- text mining
- action selection
- question answering
- natural language understanding
- reinforcement learning problems
- named entities
- optimization algorithm
- artificial intelligence
- wordnet
- dynamic programming
- machine translation
- markov decision processes
- policy evaluation
- markov decision process
- state and action spaces
- markov decision problems
- state space
- learning algorithm
- long run
- computational linguistics
- control policies
- partially observable domains
- expert systems
- sentiment analysis
- software components
- continuous state spaces
- named entity recognition
- policy gradient
- database
- rl algorithms
- average reward
- linguistic knowledge
- partially observable markov decision processes
- natural language generation
- reinforcement learning algorithms
- partially observable