User-Friendly Automatic Transcription of Low-Resource Languages: Plugging ESPnet into Elpis

  • Oliver Adams AtoszData
  • Benjamin Galliot Langues et Civilisations à Tradition Orale (LACITO), CNRS-Sorbonne Nouvelle
  • Guillaume Wisniewski Université Paris, Laboratoire de Linguistique Formelle (LLF), CNRS
  • Nicholas Lambourne The University of Queensland and ARC Centre of Excellence for the Dynamics of Language (CoEDL)
  • Ben Foley The University of Queensland and ARC Centre of Excellence for the Dynamics of Language (CoEDL)
  • Rahasya Sanders-Dwyer The University of Queensland and ARC Centre of Excellence for the Dynamics of Language (CoEDL)
  • Janet Wiles The University of Queensland and ARC Centre of Excellence for the Dynamics of Language (CoEDL)
  • Alexis Michaud Langues et Civilisations à Tradition Orale (LACITO), CNRS-Sorbonne Nouvelle
  • Séverine Guillaume Langues et Civilisations à Tradition Orale (LACITO), CNRS-Sorbonne Nouvelle
  • Laurent Besacier Laboratoire d’Informatique de Grenoble (LIG), CNRS-Université Grenoble Alpes
  • Christopher Cox University of Alberta
  • Katya Aplonova Langage, Langues et Civilisation d’Afrique (LLACAN), CNRS-INALCO
  • Guillaume Jacques Centre de Recherches Linguistiques sur l’Asie Orientale (CRLAO), CNRS-EHESS
  • Nathan Hill School of Oriental and African Studies, University of London

Abstract

This paper reports on progress integrating the speech recognition toolkit ESPnet into Elpis, a web front-end originally designed to provide access to the Kaldi automatic speech recognition toolkit. The goal of this work is to make end-to-end speech recognition models available to language workers via a user-friendly graphical interface. Encouraging results are reported on (i) development of an ESPnet recipe for use in Elpis, with preliminary results on data sets previously used for training acoustic models with the Persephone toolkit along with a new data set that had not previously been used in speech recognition, and (ii) incorporating ESPnet into Elpis along with UI enhancements and a CUDA-supported Dockerfile.

Published
2021-03-02