Showing papers on "Shallow parsing published in 1998"
Dissertation•
06 Feb 1998
TL;DR: In this paper, a relaxation labelling algorithm is applied to NLP disambiguation, where language is modelled through context constraint inspired on Constraint Grammars.
Abstract: The thesis describes the application of the relaxation labelling algorithm to NLP disambiguation. Language is modelled through context constraint inspired on Constraint Grammars. The constraints enable the use of a real value statind "compatibility". The technique is applied to POS tagging, Shallow Parsing and Word Sense Disambigation. Experiments and results are reported. The proposed approach enables the use of multi-feature constraint models, the simultaneous resolution of several NL disambiguation tasks, and the collaboration of linguistic and statistical models.
62 citations
Proceedings Article•
01 Jan 1998TL;DR: The work carried out for Italian which combines finite state parsing techniques (chunking) with analogy-based lexical acquisition with results of acquisition experiments are illustrated and evaluated.
Abstract: SPARKLE is a European project aimed at the development of robust and portable software tools for shallow parsing of unrestricted texts and for automatic induction of linguistic knowledge from corpora. In this paper we illustrate the work carried out for Italian which combines finite state parsing techniques (chunking) with analogy-based lexical acquisition. The system presented in the paper extracts predicate subcategorization frames from actual instantiations of their use in a comparatively small corpus (of about 1 million word tokens) in unsupervised mode. Results of acquisition experiments are illustrated and evaluated.
12 citations
Posted Content•
TL;DR: The thesis describes the application of the relaxation labelling algorithm to NLP disambiguation, which enables the use of multi-feature constraint models, the simultaneous resolution of several NL disambigsuation tasks, and the collaboration of linguistic and statistical models.
Abstract: The thesis describes the application of the relaxation labelling algorithm to NLP disambiguation. Language is modelled through context constraint inspired on Constraint Grammars. The constraints enable the use of a real value statind "compatibility". The technique is applied to POS tagging, Shallow Parsing and Word Sense Disambigation. Experiments and results are reported. The proposed approach enables the use of multi-feature constraint models, the simultaneous resolution of several NL disambiguation tasks, and the collaboration of linguistic and statistical models.
7 citations
01 Mar 1998
TL;DR: I n -fliis p ap er, a p u re ly log ical approach to p a rt-o f-sp eec h tag g in g an d sh a llo w p a rs in g is exp lo red .
Abstract: I n -fliis p ap er, a p u re ly log ical approach to p a rt-o f-sp eec h tag g in g an d sh a llo w p a rs in g is exp lo red . I t h as a lo t in co m m o n w ith red u c tio n is t p a rs in g s tra teg ies such as th o se em p lo y ed in C o n stra in t G ram m ar (K arlsso n e t al. 1994) an d F in ite -S ta te In te rsec tio n G ram m ar (K o sk en n iem i 1990), b u t ru les are fo rm u la ted en tire ly in log ic , an d a m o d e l g en era tio n th eo rem p ro v e r is u sed fo r p a rt-o f-sp eech ta g g in g and parsing .
6 citations