For each of the following sample we assume:
A sample from linked-extractions.tsv dataset:
1) Kemnitz is a municipality in the Ostvorpommern district
2) Shua is a village in the Astara Rayon of Azerbaijan
partOf(Shua, Astara Rayon) ^ Village(Shua)
3) Mie Prefecture forms the eastern part of the Kii Peninsula
partOf(Mie Prefecture, Kii Peninsula)
4) Mumps vaccine is a part of the MMR vaccine
partOf(Mumps vaccine, MMR vaccine)
5) Euramerica was a part of Laurasia
partOf(Euramerica, Laurasia) ??
6) Galactose is part of lactose
Dihydroergotamine is used to treat migraine headaches
What to do?
- Learn (identify) and classify the semantic relations between the Linked Entities, with emphasis on Partonomy detection. then (2nd goal)
- Generalize (subsume) an Entity “instance” in a relation to its (ontological) concept. e.g.
- See Relation Detection and Characterization (RDC)
- Semantic Web and Big Data - opportunity and challenges, See slides number 30 & 31
- Moving beyond sameAs with PLATO: Partonomy detection for Linked Data
- Learning of Semantic Relations between Ontology Concepts using Statistical Techniques, see ppt
- Unsupervised Learning of Semantic Relations between Concepts of a Molecular Biology Ontology
- Ontology Learning, Semantic Patterns, Semantic Relation Classification, Knowledge Representation, Semantic Web and Linked Data.