Hierarchical Pre-training for Sequence Labelling in Spoken Dialog

Emile Chapuis (1, 2) , Pierre Colombo (1, 2) , Matteo Manica (3) , Matthieu Labeau (1, 2) , Chloé Clavel (1, 2)

1 S2A - Signal, Statistique et Apprentissage
2 IDS - Département Images, Données, Signal
3 IBM France Lab [Biot]

Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a key component of spoken dialog systems. In this work, we propose a new approach to learn generic representations adapted to spoken dialog, which we evaluate on a new benchmark we call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE). SILICONE is model-agnostic and contains 10 different datasets of various sizes. We obtain our representations with a hierarchical encoder based on transformer architectures, for which we extend two well-known pre-training objectives. Pre-training is performed on OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We demonstrate how hierarchical encoders achieve competitive results with consistently fewer parameters compared to state-of-the-art models and we show their importance for both pre-training and fine-tuning.

Domaines: Informatique [cs]

Emile Chapuis, Pierre Colombo, Matteo Manica, Matthieu Labeau, Chloé Clavel. Hierarchical Pre-training for Sequence Labelling in Spoken Dialog. Findings of the Association for Computational Linguistics: EMNLP 2020, Nov 2020, Online, France. pp.2636-2648, ⟨10.18653/v1/2020.findings-emnlp.239⟩. ⟨hal-03134851⟩