DFKI-LT - Methods and Applications for Relation Detection
Methods and Applications for Relation Detection
1 Proceedings of the Third IEEE International Conference on Natural Language Processing and Knowledge Engineering,
The detection of relation instances is a central functionality for the extraction of structured information from unstructured textual data and for gradually turning texts into semi-structured information. Wel provide examples for the classes of approaches to relation extraction and summarize their respective advantages and disad vantages. We will argue that different relation detection tasks require different methods or even different combinations of methods. One empirically promising and theoretically attractive line of research is the learning of extraction rules from seeds.
We explain several bootstrapping methods, most of them starting with patterns as seeds and some with event seeds. We will also briefly describe our own approach of bootstrapping (Xu et al. 2007), an extension of Xu et al. (2006). In this approach, learning starts from a small set of n-ary relation instances as "seeds" in order to auto ma ti cally learn pattern rules from parsed data, which then can extract new instances of the n-ary relation and its projections. We then present a theory of the suitability of learning via bootstrapping with respect to tasks and data properties.