Rewarding Smatch: Transition-Based AMR Parsing with Reinforcement Learning.
Tahira NaseemAbhishek ShahHui WanRadu FlorianSalim RoukosMiguel BallesterosPublished in: ACL (1) (2019)
Keyphrases
- reinforcement learning
- transition model
- natural language
- state space
- natural language processing
- reinforcement learning algorithms
- markov decision processes
- function approximation
- unsupervised learning
- learning algorithm
- model free
- temporal difference
- natural language parsing
- intermediate representation
- context free grammars
- optimal policy
- transfer learning
- temporal difference learning
- dependency parser
- syntactic analysis
- autonomous learning
- dynamic programming
- robotic control
- machine learning
- language understanding
- n gram
- word order
- phrase structure
- speech understanding