Projects

VTS

With Prof. Jacques Moeschler (U. of Geneva), Prof. Genoveva Puskas (U. of Geneva), Dr. Joanna Blochowiak (CNRS, Lyon), Dr. Juan Sun (Sun Yat-sen University)

Funded by the Swiss SNSF

Verbal Tenses and Subjectivity in a cognitive pragmatics perspective

The research carried in the VTS project follows four main research lines: (i) the pragmatics of verbal tenses and their relation to subjectivity in French, (ii) the link between subjectivity and verbal tenses, precisely the Historical Present and its difference with the Free Indirect Discourse (FID); (iii) the linguistic cues of speaker’s subjectivity, and (iv) the expression of temporal reference in two typologically different languages: Mandarin on the one hand and French and English on the other hand.

Processing temporal relations

With Dr. Hannah Rohde

Funded by the Swiss SNSF

Pieces of discourse such as The plane landed, (then) passengers got off and The plane landed unexpectedly because it had a technical problem show that discourse connectives, generally used to explicitly express a discourse relation, sometimes are necessary and other times they become redundant. Scholars to date have not attended to the divergent behaviour of the various types of discourse relations and their explicit expression through discourse connectives. Assessing this issue is necessary for having an accurate understanding of how humans process pieces of discourse and how does linguistically encoded information pilot this process.

MODERN

With Prof. Jacques Moeschler (U. of Geneva), Prof. Andrei Popescu-Belis (IICT informatique et télécommunication HEIG-VD), Prof. Martin Volk (Institut für Computerlinguistik Universität Zürich), Prof. Ted Sanders (Utrecht Institute of Linguistics OTS University of Utrecht), Prof. Sandrine Zufferey (U. of Bern)

Funded by the Swiss SNSF

Modeling Discourse Entities and Relations for Coherent Machine Translation

The goal of MODERN is to model and automatically detect textual dependencies across sentences, and to study their integration within machine translation systems, with the aim of demonstrating improvement in translation quality, with a specific focus on the interplay between referring expressions and discourse relations in four languages: English, French, German and Dutch.

COMTIS

With Prof. Jacques Moeschler (U. of Geneva), Prof. Andrei Popescu-Belis (IICT informatique et télécommunication HEIG-VD), Prof. Sandrine Zufferey (U. of Bern), Prof. Paola Merlo (U. of Geneva), Dr. James Henderson (IDIAP Research Institute), Dr. Bruno Cartoni (Google), Dr. Thomas Meyer (Google)

Funded by the Swiss SNSF

Improving the Coherence of Machine Translation Output by Modeling Intersentential Relations

The objective of the COMTIS SNSF Sinergia project was to use insights from linguistic modeling and corpus linguistics in order to build computational models of discourse-level phenomena and to combine them with statistical machine translation systems, thus improving the quality of translated texts. The COMTIS researchers have advanced the state of the art in their fields and with respect to the overall objective, thanks to the close collaboration of all partners, also reflected in the quality and number of joint publications. More specifically, we have proposed multilingual models of discourse connectives and verb tenses, strongly grounded in empirical evidence from parallel corpora, mainly in English and French, but also German, Italian, and Arabic. These models have generated features which served to implement automatic labeling systems for discourse connectives, verbs, and pronouns, which were further combined, in several ways, with state-of-the-art statistical MT systems and with innovative tree-based decoding algorithms. This has led to demonstrable improvements of the MT output, as assessed by humans but also by an automatic reference-based metric which COMTIS proposed and validated.

%d bloggers like this: