A Within-Frame Ontological Extension on FrameNet: Application in Predicate Chain Analysis and Question Answering
An ontological extension on the frames in FrameNet is presented in this paper. The general conceptual relations between frame elements, in conjunction with existing characteristics of this lexical resource, suggest more sophisticated semantic analysis of
- PDF / 311,557 Bytes
- 11 Pages / 430 x 660 pts Page_size
- 54 Downloads / 199 Views
Centre for Informatics and Applied Optimization School of Information Technology and Mathematical Sciences, University of Ballarat PO Box 663, Ballarat VIC 3353, Australia [email protected], {j.yearwood, l.ma}@ballarat.edu.au
Abstract. An ontological extension on the frames in FrameNet is presented in this paper. The general conceptual relations between frame elements, in conjunction with existing characteristics of this lexical resource, suggest more sophisticated semantic analysis of lexical chains (e.g. predicate chains) exploited in many text understanding applications. In particular, we have investigated its benefit for meaning-aware question answering when combined with an inference strategy. The proposed knowledge representation mechanism on the frame elements of FrameNet has been shown to have an impact on answering natural language questions on the basis of our case analysis. Keywords: FrameNet, Ontological Relation, Question Answering.
1 Introduction In many cases, most knowledge underlying a lexical resource can be extracted when the knowledge is represented in a network of relations between the constituents of a resource. In such a network, both the nodes (constituents) and the edges (relations) are knowledge-bearing. Such relations formulate the interactions happening between constituents in a more sophisticated fashion. Understanding these interactions can help in text understanding applications like predicate chain analysis. By predicate chains we refer to a specific type of lexical chains where relations between the lexical units (lexemes) are realized by the concept of the predicates. With such a definition, predicate chains differ from the three main types of lexical chains with WordNet-based relations (i.e. extra-strong, strong, and regular). We have identified a mapping strategy which transforms predicative relations to some more general ontological relations that can be useful in text understanding applications. Question answering (QA) is one of the areas where predicate chain analysis can help. On some occasions, the answer entity is hidden behind a chain of predicative relations or it can be justified only after pursuing such relations in the answer-bearing M.A. Orgun and J. Thornton (Eds.): AI 2007, LNAI 4830, pp. 404–414, 2007. © Springer-Verlag Berlin Heidelberg 2007
A Within-Frame Ontological Extension on FrameNet
405
texts. For instance, the sentence “they first met on the bridge which was later named as Johnston and was one of the adventures in the area of Mount Clear…” for the question “who first met in Mount Clear?” contains the actual answer that can be reached when following the relations between the entities “bridge”, “Johnston”, and “Mount Clear”. Accessing such an answer entity requires that an automated system be capable of semantically resolving the constituents and connecting them all together in an inference mechanism to identify the right answer. We have been investigating FrameNet [1] as a lexical resource to achieve better answer extraction. To address
Data Loading...
