Bich-Liên Doan is a Professor (HDR) at CentraleSupélec and a member of LISN, in the A&O team (Apprentissage et Optimisation — Learning and Optimization).

Her research mainly concerns information retrieval, in particular the use of machine learning and natural language processing algorithms in recommender systems. More specifically:

• information representation (NLP and ontologies) and unsupervised learning models, including deep neural networks, in order to disambiguate queries and improve precision in recommender systems,
• taking context into account when defining new paradigms for modeling information retrieval (query reformulation, task context, dynamics of the information retrieval system),
• analysis of non-classical models to improve the performance of information retrieval systems (unification of probabilistic and logical models, vector models extended to Hilbert spaces).

She has published 30 articles in information retrieval and recommender systems conferences and journals, more recently using deep neural network techniques.