LST day

On 27 and 28 November, the LST team will be holding its research days. This is an opportunity for young and more experienced researchers to present their research topics. In addition, this year we are delighted to welcome two external speakers, Mathieu Fontaine from Télécom Paris and Adrien Bardet from Atlantique Traduction, who will each launch a day with a seminar.

This year’s presentations will be organized according to four axes:

• axis 1: lexical embedding, interpretability, graphs, lexicon, frugal, interdisciplinary

• axis 2: speaker, explicability, evaluation, interruption, superimposed speech, corpus design, speaker, diarisation, wood dendrochronology, automatic segmentation, microphone antennas, statistical models, assisted learning, active learning, TTS

• axis 3: speech understanding, opinion detection, automatic language processing, expressive speech, synthesis, characterisation, explicability, voice identity, prosody, pronunciation, deep fake speech, production and perception of speech and language, phonetic variation, control, speech synthesis, voice conversion.

• axis 4: speech translation, speech recognition, representation extraction, under-resourced language, cross-lingual representation/p>