Maximum Margin Reward Networks for Learning from Explicit and Implicit Supervision.
Haoruo PengMing-Wei ChangWen-tau YihPublished in: EMNLP (2017)
Keyphrases
- maximum margin
- structured prediction
- structured output
- learning algorithm
- learning process
- reinforcement learning
- active learning
- markov networks
- support vector
- pattern classification
- feature selection
- data sets
- probabilistic model
- support vector machine
- upper bound
- supervised learning
- image classification
- learning tasks
- principal components
- hidden variables