ASR-Generated Text for Language Model Pre-training Applied to Speech Tasks

Valentin Pelloin (1) , Franck Dary (2) , Nicolas Hervé (3) , Benoît Favre (2) , Nathalie Camelin (1) , Antoine Laurent (1) , Laurent Besacier (4)

1 LIUM - Laboratoire d'Informatique de l'Université du Mans
2 LIS - Laboratoire d'Informatique et des Systèmes (LIS) (Marseille, Toulon)
3 INA - Institut National de l'Audiovisuel
4 Naver Labs Europe [Meylan]

We aim at improving spoken language modeling (LM) using very large amount of automatically transcribed speech. We leverage the INA (French National Audiovisual Institute 1) collection and obtain 19GB of text after applying ASR on 350,000 hours of diverse TV shows. From this, spoken language models are trained either by fine-tuning an existing LM (FlauBERT 2) or through training a LM from scratch. New models (FlauBERT-Oral) are shared with the community and evaluated for 3 downstream tasks: spoken language understanding, classification of TV shows and speech syntactic parsing. Results show that FlauBERT-Oral can be beneficial compared to its initial FlauBERT version demonstrating that, despite its inherent noisy nature, ASR-generated text can be used to build spoken language models.

Domaines: Informatique [cs]; Intelligence artificielle [cs.AI]; Apprentissage [cs.LG]; Traitement du texte et du document; Traitement du signal et de l'image [eess.SP]

Valentin Pelloin, Franck Dary, Nicolas Hervé, Benoît Favre, Nathalie Camelin, et al.. ASR-Generated Text for Language Model Pre-training Applied to Speech Tasks. Interspeech 2022, Sep 2022, Incheon, South Korea. ⟨hal-03770506⟩