IRIT - MELODI

Knowledge modeling and natural language analysis


Our team, at a glance

(~ 14 researchers)

Our team embraces various theoretical and applicative aspects of the duality language/knowledge. We deal with (1) natural language analysis and its formalization and with (2) knowledge modeling.We are renown worldwide for ontology (representation of the world) and text skills, for lexical semantic (word, text, dialogs), structuring knowledge through graphs or logic, automatical learning of structured representation, building ontologies from texts and allowing their evolution, semantic querying in large knowledge basis, formal checking in knowledge basis.

Applications

  • Textual structure interpretation
  • Machine learning and distributional approaches (stats measure of co-occurring words) for information extraction
  • Ontologies and information retrieval : An application to automotive diagnostics (ACTIA)
  • Collaborative decision-making and decision support systems for enhancing operations management in industrial environments (SIMTECH SIMULATION)
  • Ontologies for explaining the data coming from sensors
  • Measuring the effect of discourse structure on sentiment analysis (SYNAPSE)
  • Textual structure interpretation
  • Ontology co-construction with an adaptive multi-agent system : Principles and case-study (UPETEC)