Graph-based probabilistic induction and learning for database systems
Pohang University of Science and Technology (POSTECH)
Oracle Principal Investigator
Hassan Chafi, Vice President, Research and Advanced Development
Sungpack Hong, Director, Research And Advanced Development
Graph-based probabilistic induction and learning is an emerging topic that learns probabilistic recursive rules from probabilistic graph data. We can derive new probabilistic facts from these rules.
In the first year of this project, we investigate breakthrough technologies in inductive learning for graph data. First, we study how to incorporate induction in deriving regular path queries for graph data with positive and negative examples. This has many real applications, such as knowledge bases and graph mining. Second, we study how to extend the power of the regular path expression to support generalized subgraph isomorphism queries containing regular paths. This technique can be directly applied to derive inference rules from RDF data.