Pro3Gres: A low-complexity, broad-coverage statistical Dependency Parser for English

The Pro3Gres Parser: Overview

Pro3Gres stands for PRObability-based, PROlog-implemented Parser for RObust Grammatical Relation Extraction System. It is a fast, broad-coverage, deep-syntactic parsing system. It is a flexible and perspicuous hybrid parser using easily editable hand-written rules and statistical lexicalization from the Penn Treebank. Its performance is state-of-the-art or almost state-of-the-art. Its statistical model is based on the decisions that (a human or a machine) parser has to take during the parsing process.

It is a Dependency Grammar system, inspired by the works of Tesnère and Lexcial-Functional Grammar (LFG) [Schneider et al. 2004b].

Potential applications of Pro3Gres, some of which are already being explored, include:

• Robust, broad-coverage fast parsing of unrestricted text [most articles, esp. Schneider et al. 2004a]
• Deep-linguistic parsing, including long-distance dependencies [esp. Schneider 2003]
• Question Answering [Rinaldi et al. 2004a]
• Relation Extraction and building a domain-specific Knowledge Database [Rinaldi et al. 2004b]
• Parsing BioMedical Texts [Rinaldi et al. 2004b]: additional link: Why is Pro3Gres suitable for Parsing Medical Texts ?
• High-Precision Parsing [link to follow soon]
• Helping annotators to build Treebanks [link to follow soon]
• Psycholinguistically adequate parsing, testing of psycholing. hypothesis [link to follow soon]
• Exploring hybridness [Schneider et al. 2004a]
• Exploring combinations of shallow and deep-processing [Schneider 2004]
• Framenet-based Theta Role parsing
• Domain-specific customized parsing

Selected Publications and Presentations

• [Schneider et al. 2004b] Gerold Schneider, Fabio Rinaldi, James Dowdall, 2004. Fast, deep-linguistic statistical dependency parsing. Recent advances in DG workshop, Coling 2004, Geneva.
• [Schneider et al. 2004a] Gerold Schneider, James Dowdall, Fabio Rinaldi, 2004. A robust and hyrid deep-linguistic theory applied to deep-linguistic parsing. Robust Methods in the analysis of NL data (ROMAND) workshop, Coling 2004, Geneva.
• [Schneider 2004] Gerold Schneider, 2004. Combining Shallow and Deep Processing for a Robust, Fast, Deep-Linguistic Dependency Parser. European Summer School in Logic, Language and INformation (ESSLLI) 2004, Nancy.
• [Rinaldi et al. 2004b] Fabio Rinaldi, Gerold Schneider, James Dowdall, Christos Andronis, Andreas Persidis, Ourania Konstanti, 2004. Mining relations in the GENIA corpus.In: Proceedings of the second workshop on Data Mining and Text Mining for Bioinformatics. Pisa, Italy.
• [Rinaldi et al. 2004a] Fabio Rinaldi, James Dowdall, Gerold Schneider, Andreas Persidis, 2004. Answering Questions in the Genomics Domains. In: ACL 2004 Workshop on QA in restricted domains, Barcelona, Spain.
• TaCoS 2004 in Bonn presentation slides
• [Schneider 2003] Gerold Schneider, 2003. Extracting and Using Trace-Free Functional Dependencies from the Penn Treebank to Reduce Parsing Complexity. In: Joakim Nivre and Erhard Hinrichs, eds.: Proceedings of Treebanks and Linguistic Theories (TLT) 2003. Växjö University Press, Series: Mathematical Modelling in Physics, Engineering and Cognitive Sciences, v. 9, Växjö, Sweden: 153-164.
• Link to my Linguistik Online 2003 articles
• Dependency Grammar Lecture Material, mostly in English, partly in German
• the slides of my NAACL/HLT 2003 presentation
• Geneva Doctoral Candidate Presentation
• Zurich Doctoral Candidate Presentation
• [Henderson et al. 2002a] J.Henderson, P.Merlo, I.Petroff, and G.Schneider. Using Syntactic Analysis to Increase Efficiency in Visualizing Text Collections. In Proceedings of the 19th International Conference on Computational Linguistics (COLING 2002), pages 335-341, Taipei, Taiwan, 2002.
• [Henderson et al. 2002b] J.Henderson, P.Merlo, I.Petroff, and G.Schneider. Using NLP to Efficiently Visualize Text Collections with SOMs. In Proceedings of the 3rd International Workshop on Natural Language and Information Systems (NLIS 2002), Aix-en-Provence, France, 2002.
• more of my publications.

Gerold Schneider
Institute of Computational Linguistics
University of Zurich        
Binzmühlestrasse 14      
CH-8050 Zurich             
                                 
Email: gschneid AT cl.unizh.ch