Training with Exploration Improves a Greedy Stack-LSTM Parser.
Miguel BallesterosYoav GoldbergChris DyerNoah A. SmithPublished in: CoRR (2016)
Keyphrases
- greedy algorithm
- feature selection
- natural language processing
- natural language
- search algorithm
- supervised learning
- training examples
- test set
- training process
- training set
- dynamic programming
- support vector
- artificial intelligence
- training phase
- search space
- worst case
- online learning
- recurrent neural networks
- dependency parsing