Parsing noun phrases with interpreted regular tree grammars
Several common tasks in natural language processing (NLP) involve graph transformation, in particular those that handle syntactic trees, dependency structures such as Universal Dependencies (UD) [1], or semantic graphs such as AMR [2] and 4lang [3]. Interpreted Regular Tree Grammars (IRTGs) [4] enco...
Elmentve itt :
| Szerzők: | |
|---|---|
| Testületi szerző: | |
| Dokumentumtípus: | Könyv része |
| Megjelent: |
2019
|
| Sorozat: | Magyar Számítógépes Nyelvészeti Konferencia
15 |
| Kulcsszavak: | Nyelvészet - számítógép alkalmazása |
| Online Access: | http://acta.bibl.u-szeged.hu/59094 |
| Tartalmi kivonat: | Several common tasks in natural language processing (NLP) involve graph transformation, in particular those that handle syntactic trees, dependency structures such as Universal Dependencies (UD) [1], or semantic graphs such as AMR [2] and 4lang [3]. Interpreted Regular Tree Grammars (IRTGs) [4] encode the correspondence between sets of such structures and have in recent years been used to perform both syntactic and semantic parsing. In this paper we introduce our tool that is capable of automatic IRTG generation. Our IRTG covers 83% of noun phrases (NPs) from the Wall Street Journal section of the Penn Treebank and a pilot experiment had also been made for retrieving surface realizations from UD graphs using independent data. We also describe this generated IRTG which allows for simultaneous generation of structures of various types and can be used for semantic parsing, generation, and semanticsdriven translation. |
|---|---|
| Terjedelem/Fizikai jellemzők: | 301-313 |
| ISBN: | 978-963-315-393-2 |