Department - Natural Language and Knowledge Extraction
PROFILE

Natural Language and Knowledge Extraction

  
 

The Department conducts basic and applied research in the field of Natural Language Processing. Its main aim is to design computational models for natural language recognition and "understanding". In particular, the Natural Language and Knowledge Extraction Department designs, implements and integrates human language technologies in systems and applications dealing with structured data as well as with unstructured data processing. Applications include language-aware information retrieval and extraction, natural language interfaces, text mining, knowledge management, etc. To achieve its goals, the Department builds and deploys multi-level language resources necessary in the technology production process.

Natural Language and Knowledge Extraction Department research and development efforts aim at developing technology amenable to multilingual and multimedia information processing. The core of the approach adopted consists in developing, improving and valorising methods and techniques in the areas of artificial intelligence, machine learning, statistical and rule-based processing as well as hybrid combinations in an attempt to propose solutions that are flexible and adaptive to a range of linguistic phenomena and applications.

Multilinguality is catered for by resorting to processing of parallel and comparable text corpora and elicitation of multilevel translation equivalents. Multimediality is handled by integrating conversion technologies, able to convert speech and image modalities to text via automatic speech recognition and image processing.

Tools are designed and developed for:

  • efficient structural and grammatical text annotation
  • shallow parsing
  • named entity recognition
  • term extraction
  • co-reference resolution
  • event recognition and their interrelations
  • semantic-web oriented information extraction
  • text classification
  • word sense disambiguation
  • multilevel text corpora matching
  • multilevel parallel text alignment
  • automatic elicitation of lexical data pertinent to monolingual and multilingual applications
  • intelligent translation memory applications

The Department is active in the framework of competitive national and European Union projects.

 
 

Current Projects

Services

Products

 
Team members: