DeepPath: A Reinforcement Learning Method for Knowledge Graph Reasoning.
Wenhan XiongThien HoangWilliam Yang WangPublished in: EMNLP (2017)
Keyphrases
- high precision
- computational cost
- similarity measure
- high accuracy
- domain knowledge
- pairwise
- knowledge base
- tree structure
- significant improvement
- prior knowledge
- reinforcement learning
- preprocessing
- random walk
- detection method
- neural network
- graph representation
- computational complexity
- video sequences
- cost function
- multi agent
- objective function
- knn
- knowledge based systems
- segmentation algorithm
- image segmentation