HackRL: Reinforcement learning with hierarchical attention for cross-graph knowledge fusion and collaborative reasoning.
Linyao YangXiao WangYuxin DaiKejun XinXiaolong ZhengWeiping DingJun ZhangFei-Yue WangPublished in: Knowl. Based Syst. (2021)
Keyphrases
- reinforcement learning
- knowledge representation
- knowledge base
- knowledge sharing
- general knowledge
- deductive reasoning
- reasoning process
- domain knowledge
- incomplete knowledge
- formal models
- procedural knowledge
- knowledge intensive
- knowledge based systems
- knowledge acquisition
- knowledge management
- probabilistic knowledge
- expert systems
- uncertain information
- information fusion
- knowledge structures
- model based reasoning
- reasoning mechanism
- commonsense knowledge
- directed graph
- solving problems
- graph representation
- graph mining
- graph matching
- data fusion
- structured data
- random walk
- data mining techniques
- state space
- knowledge discovery
- prior knowledge
- multi agent
- bayesian networks