TALN - Automatic Natural Language
The mass of language data now available makes it possible to implement robust, language-independent techniques. However, data quantity does not always mean quality, robustness and analytical finesse. The TALN team tries to reconcile these two antagonistic aspects by proposing robust text analysis methods that can be adapted to the diversity of written language data expressed on new communication media such as blogs, social networks, forums, coupling with other media or expressing themselves in different languages. TALN work is multidisciplinary in nature, at the heart of the data, in interaction with the human and social sciences (linguistics, terminology, translation, information sciences) and with other computer science themes such as learning, data mining, signal recognition (speech, gesture), information retrieval.
The team's work focuses on written language analysis and falls under two main research themes :