Prof. Dr. Peter Hellwig's expertise
- Work on grammar theory and inferential semantics for more than fourty
years, which resulted in the development of Dependency Unification Grammar
(DUG), a lexicalistic approach to the description and automated processing
of languages.
- Work on parsing theory, which has been summarized in a comprehensive survey
of tasks and choices associated with the construction of parsers, published
in the De Gruyter Handbook of Computational Linguistics.
- Development of an innovative lexicon-based chart parser, based on the
slot-and-filler principle.
- Development of a comprehensive theory of cohesion and coherence of texts and
the connection between sentence structure and text structure. This theory has
a great potential for text summarizing.
- Experience in software development, starting with the construction
of PLAIN (Programs for Language Analysis and Inference) in the seventies. PLAIN is a collection of programs which support the writing of DUG
grammars and include a parsing and inferencing module. Recently, the system
has been re-implemented in C under UNIX with a client-server architecture
and the potential for massive parallel processing.
- Experience in various applications. In the ESPRIT project "Translator's
Workbench" (TWB), the computational linguists of the University of
Heidelberg developed a module for grammar checking and correction. A lemmatizer
and morpho-syntactic tagger has been implemented as well as a morphological
and syntactic training function for CALL applications. The Heidelberg parser
has been used in the LRE-project "Selecting Information from Texts"
(SIFT), which aims at retrieving passages in texts that match user queries.
In the LE-project "Multilingual Authoring of Business Letters" (MABLe),
Heidelberg was charged with the generation of variable portions within
the letters.
- Experience in developing broad coverage lingware such as the German
lexicon of 90,000 items which was ordered by an industrial partner. Large
scale morphological and syntactic lexicons of German, French and English
have been developed in the projects TWB and SIFT.
- Engagement in the standardization process. Professor Hellwig co-authored
the documents on applications and on the feasibility of standards for syntactic
descriptions in the framework of the EUROTRA-7 Study on Re-usable Lexical
and Terminological Resources. He was a member of the Working Group on the
Computational Lexicon in the "Expert Advisory Group for Language Engineering
Standards" (EAGLES).
- Experience in lexical semantics, originating from an INTAS project (CALLEX) about lexical
functions together with the Computational Linguistics Laboratory in Moscow. At presesent,
new lexical relationships are explored to be included in a wordnet-like database.
- Interest in the transformation between graphic and linguistic representations, e.g.
between images and their descriptions and between maps and oral navigation instructions.
- Experience in the development of curricula and courses in computational linguistics since many years.
The installation of a new Bachelor- and a Masters degree in Computational Linguistics
according to international standards has been put through in Heidelberg.
|
