Path Reasoning over Knowledge Graph: A Multi-agent and Reinforcement Learning Based Method.
Zixuan LiXiaolong JinSaiping GuanYuanzhuo WangXueqi ChengPublished in: ICDM Workshops (2018)
Keyphrases
- multi agent
- reinforcement learning
- prior knowledge
- detection method
- clustering method
- knowledge base
- dynamic programming
- reasoning process
- knowledge representation
- domain knowledge
- significant improvement
- similarity measure
- cost function
- state space
- support vector machine
- high accuracy
- pairwise
- computational complexity
- structured data
- graph matching