| Abstract |
Open Information Extraction is the task of extracting triples (of relations and the entities they relate) from free text, without any external sources of knowledge, such as, e.g., ontologies or knowledge bases. Disambiguating these entities is an additional task. Which surface forms refer to the same entity? Which relations are semantically similar? We aim to do this disambiguation step in an 'open' fashion, so unsupervised and without any external sources of information.
In this talk I will discuss some ideas I have and I hope to receive helpful feedback.
This is very much ongoing research, so no tables with smashing results (yet…). |