Duy Dinh

IRIT laboratory - Paul Sabatier University
Toulouse, France

   
 
  Duy Dinh PhD candidate
Laboratory: IRIT
Team: Systèmes d'Information Généralisés
Localization: IRIT1 / Level 4, Office: 406
Office number: (+33)5 61 55 73 91
Contact: Duy[dot]Dinh[at]irit[dot]fr
118 route de Narbonne
31062 Toulouse, Cedex 04
“Knowledge is of no value unless you put it into practice.”

Heber J. Grant

 

I am a Ph.D candidate, ATER at the Paul Sabatier University. I am working in the SIG team under the supervision of Dr. Lynda Tamine-Lechani, Assistant professor at the Paul Sabatier University.

I am interested in various techniques and approaches to information retrieval (IR), especially conceptual or semantic indexing and retrieval methods in the biomedical domain. My ongoing thesis focuses on several Natural Language Processing (NLP) techniques such as Word/Term Sense Disambiguation, String Matching, Concept extraction or concept mapping, Term mismatch resolution, ... to improve the IR effectiveness.

Research interests:

  • Information Retrieval (IR): Weighting models (Term/Document relevance estimation), Evaluation,
  • Biomedical IR: Genomics IR, Medical Records retrieval, Termino-ontological IR,
  • Information extraction: Approximate dictionary lookup, statistical concept extraction,
  • Natural Language Processing: Etimology, Part-Of-Speech tagging, Word/Term sense disambiguation,
  • Term mismatch resolution: Document and/or Query Expansion, Document and/or Query Reduction,
  • Controlled terminologies: biomedical thesauri, ontologies, dictionaries,
  • Machine learning: supervised (TiMBL, SVM...), unsupervised (corpus-based).

A list of my research publications may be a better guide (also from IRIT database or DBLP).

 

Open Source Software:

  • Information Retrieval Toolkit (IRToolkit) is an attempt to develop a general Information Retrieval platform integrating several open source search engines (e.g., Terrier, Lucene, Lemur, etc.) into a single (GUI-based) application. Furthermore, it offers a capability to compare the performance (in terms of precision, recall, index size, search response time and so on) between several open source IR platforms.

  • Concept eXtractor (cxtractor) is a generic concept extractor, which aims at integrating state-of-the-art term/concept extraction methods. Furthermore, it offers a capability to compare the performance (in terms of precision, recall, F-measure) between several term/concept extraction algorithms. [bibtex]

Related conferences:

  • ECIR (European Conference on Information Retrieval)
  • AIME (Artifical Intelligence in MEdicine)
  • TREC (Text REtrieval Conference)
  • SAC (Symposium on Applied Computing)
  • NLDB (Conference on Applications of Natural Language to Information Systems)
  • CLEF (Cross-Language Evaluation Forum)