NAVER LABS Europe virtual seminars are open to the public. Please register here for your participation (Zoom event).
Date: 7th January 2025, 11:00 am (CET)
Multilingual spoken language understanding
About the speaker: Gaelle Laperrière recently defended her thesis on Spoken Language Understanding at the Laboratoire Informatique d’Avignon. Her work focused on semantic extraction with cross-lingual and cross-domain portability, processing low resource and dialectal languages. She is currently working at the LIA as a Research Engineer on Natural Language Processing.
Abstract: The primary objective of my thesis has been to leverage existing data of large resourced annotated languages for speech semantics to develop effective understanding systems in low resourced languages. In recent years, significant advances were made in the field of automatic speech translation through new approaches that converge audio and textual modalities, the latter benefiting from vast amounts of data. By visualizing spoken language understanding as a translation task from a natural source language to a conceptual target language, we considered the SAMU-XLSR speech encoder, which generates a semantically enriched language-agnostic encoding. This presentation will demonstrate the positive impact of this type of encoder in an end-to-end speech understanding neural network and highlight its linguistic and semantic encoding capabilities. It will also present our exploration of two different specializations of its semantic enrichment, aiming to direct its encoding towards the semantic domain of the French MEDIA, Italian PortMEDIA, and Tunisian TARIC-SLU tasks and therefore participating in improving state-of-the-art in portability between languages and domains for MEDIA, PortMEDIA, and TARIC-SLU datasets.