Peter Kostelník (Xolution), Martina Ivanová (Faculty of Arts, PU in Prešov, Ľ. Štúr Institute of Linguistics of the Slovak Academy of Sciences, v. v. i.) : Scenes-and-frames Semantics and its possibilities in building a knowledge database of the Slovak language
The presentation will focus on the introduction the SENSE project (Semantic Analysis of the Slovak Language), in which the Ľ. Štúr Institute of Linguistics, Slovak Academy of Sciences participates together with the Xolution company. The aim of the SENSE project is to design a knowledge database for the Slovak language which would describe how individual words or phrases are transformed into a framework representation, and the creation of software that can use these datasets to interpret the texts on which the machine has not been trained.
To achieve this purpose, the Scenes-and-frames Semantics as introduced in the FrameNet database will be applied. In the methodology of Scenes-and-frames Semantics, frame is defined as a script-like conceptual structure that describes a particular type of situation, object, or event along with its participants and properties. At the same time, FrameNet models meaning in a human-readable way, which also assumes that texts with the same meaning in different languages will be semantically interpreted in the same way, i.e., the semantic representation of the text will be the same regardless of language.
One of the tools that may help in the development of such a database are databases which have arisen from the research on valency properties as developed in the tradition of Slovak linguistics. The existing valency dictionaries of Slovak language may help to overcome one shortcoming of the FrameNet approach, namely, that the coverage of FrameNet is insufficient. It is mainly caused by the methodology of its building, which proceeds frame by frame rather than lemma by lemma. A specific feature of the valency databases of Slovak, on the other hand, is the consistent description of all lemmas of verbs and their consistent assignment to particular semantic classes (which partly have frame-like properties). The presentation will show how these methodologies can be mutually beneficial. In the presentation, practical examples of transformation of text into a framework representation through the experimental software in its pilot stage of development will be shown. The conceptual design and preliminary architecture of the semantic analyser will also be introduced. At the same time, possible pitfalls that syntactic structuring of semantic representation in Slovak may bring about will be pointed out.
Martina Ivanová is Professor of Slovak linguistics at the University of Prešov and the researcher at Ľ. Štúr Institute of Linguistics, Slovak Academy of Sciences, Bratislava (Slovakia). Her research focuses on Slovak grammar, morphematics, corpus and cognitive linguistics and teaching Slovak as a second language. She has published monographs on modal verbs and modality, cognitive grammar, verb valency, morphematics and derivational morphology. She is the co-author of five monolingual dictionaries of Slovak language (valency dictionary, dictionary of root morphemes and dictionary of word formation means). Currently, she is also the editor of monolingual dictionary of contemporary Slovak, Slovník súčasného slovenského jazyka (since 2018).
Peter Kostelník is researcher and developer at Xolution, s.r.o. His main areas of interest comprise computational linguistics mostly focused on natural language processing / understanding, semantic modelling, knowledge engineering and expert systems. He participated at several EU R&D projects, where he was commonly responsible for semantic and knowledge technologies and application of artificial intelligence methods. Peter published several journal and conference papers and participated in 2 monography books in area of cognitive sciences. He is also author of multiple natural language and knowledge processing technologies used in Xolution products.