Language Technology for Business Applications

Topics of Interest

The International Conferences on Business Information Systems (BIS) provide a forum for the dissemination of research on the design, implementation, application and improvement of computer systems for business purposes.

The aim of the special session "Language Technology for Business Applications is to present Natural Language
Processing (NLP) solutions aimed at business applications, as well as to expose NLP researchers and practitioners to the challenges encountered in the area of the deployment of language technology in real-world business applications. The session will also provide an interdisciplinary forum face-to-face exchange of new ideas and experiences, and for the discussion of future research directions and utilization
of language technology in business information systems.

We invite articles and demonstrations concerning intelligent text analysis and navigation tools, NLP-aided knowledge discovery and intelligent document management systems, as well as core linguistic processing components potentially deployable in business applications. Articles and demonstrations discussing, in the context of potential business applications, solutions and technologies within any of the following areas are in the scope of the session:

• Language Technologies in Knowledge Acquisition


• Language Technologies in Knowledge Management


• Text Classification


• Text Summarization


• (Cross-Lingual) Information Retrieval


• (Multi-Lingual) Information Extraction


• Text Mining


• Web Mining


• Textual Question Answering


• Machine and Machine-Aided Translation

Session program (23.04.2004)

Finite-State Lexical Tools

Jan Daciuk
Department of Knowledge Engineering
Gdansk University of Technology


Abstract

The paper presents three software packages containing finite-state lexical tools: two tool sets of standalone programs and support scripts – one using recognizers, and the other one using transducers, as well as a library of functions. All three packages offer similar functionality. Rather than describing the packages in details, the functions they provide are emphasized.

A General Purpose Machine Translation System for The Polish Market

Marek Labuzek
TECHLAND


Abstract

The article presents various aspects of constructing a commercial English-Polish machine translation system in the Polish market reality. Since the requirements of quick effects were laid on the project, various heuristic methods had to be employed, especially in parser construction. A combination of machine learning, standard statistical methods of tagging and hand-crafted rules produced quick effects, still leaving some possibility for constant quality increase.

SProUT: An Integrated Set of Tools for Shallow Text Processing

Jakub Piskorski
DFKI GmbH
Germany


Abstract

This paper briefly presents SProUT – a novel platform for the development of multi-lingual shallow text processing systems, developed at the Language Technology Lab of the German Research Center for Artificial Intelligence. We focus on SProUTs’ particularities concerning linguistic processing components, grammar formalism, and ongoing projects which deploy SProUT.

Finite-State Text Processing in a Speech Synthesis System

Wojciech Skut
Rhetorical Systems Ltd.
Edinburgh, Scotland


Abstract

This paper describes the architecture of the finite-state text processing components of rVoice, a text-to-speech (TTS) system developed by Rhetorical Systems. It focuses on the original motivation, the application areas of finite-state technology in TTS, typical theoretical and technical problems and the solutions found in the implementation.