Skip to main content
Springer Nature Link
Log in

Comparing NLP Solutions for the Disambiguation of French Heterophonic Homographs for End-to-End TTS Systems

  • Conference paper
  • First Online:
Speech and Computer (SPECOM 2022)

Part of the book series: Lecture Notes in Computer Science ((LNAI,volume 13721))

Included in the following conference series:

  • 1104 Accesses

Abstract

This paper presents a study on different NLP solutions for French homographs disambiguation for text-to-speech systems. Solutions are compared using a home-made corpus of 8137 sentences extracted from the Web, comprising roughly one hundred instances of each of 34 pairs of prototypical words. A disambiguation system based on per-case Linear Discriminant Analysis (LDA) classifiers using contextual word embeddings as input features achieves state-of-the-art F-scores superior to 0.96.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others

Notes

  1. 1.

    https://fr.wiktionary.org/wiki/Categorie:Homographes_non_homophones_en_francais

  2. 2.

    e.g. the root “techniqu” is only used 5 times in our audiobook database: with no additional patterns from a pronunciation lexicon, “ch” will likely be mispronounced with the post-alveolar fricative .

  3. 3.

    https://huggingface.co/gilf/french-postag-model.

References

  1. Bisani, M., Ney, H.: Joint-sequence models for grapheme-to-phoneme conversion. Speech Commun. 50(5), 434–451 (2008)

    Article  Google Scholar 

  2. Black, A.W., Lenzo, K., Pagel, V.: Issues in building general letter to sound rules. In: The Third ESCA/COCOSDA Workshop (ETRW) on Speech Synthesis. Jenolan Caves House, Blue Mountains, Australia (1998)

    Google Scholar 

  3. Bosse, M.L., Tainturier, M.J., Valdois, S.: Developmental dyslexia: the visual attention span deficit hypothesis. Cognition 104(2), 198–230 (2007)

    Article  Google Scholar 

  4. Goldman, J.P., Laenzlinger, C., Wehrli, E.: La phonétisation de plus, tous et de certains nombres: une analyse phono-syntaxique. Actes de TALN99, Cargese, Corse, pp. 165–174 (1999)

    Google Scholar 

  5. Gorman, K., Mazovetskiy, G., Nikolaev, V.: Improving homograph disambiguation with supervised machine learning. In: Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018) (2018)

    Google Scholar 

  6. Kastner, K., Santos, J.F., Bengio, Y., Courville, A.: Representation mixing for TTS synthesis. In: ICASSP 2019–2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 5906–5910. IEEE (2019)

    Google Scholar 

  7. Kumar, A.: NLP pre-trained models explained with examples (2021)

    Google Scholar 

  8. Le, H., et al.: Flaubert: unsupervised language model pre-training for French (2019). https://arxiv.org/abs/1912.05372

  9. Lenglet, M., Perrotin, O., Bailly, G.: Modélisation de la parole avec tacotron2: analyse acoustique et phonétique des plongements de caractère. In: 34\(^{e}\) Journées d’Études sur la Parole (JEP), pp. 845–854. Noirmoutier, France (2022)

    Google Scholar 

  10. Nicolis, M., Klimkov, V.: Homograph disambiguation with contextual word embeddings for TTS systems. In: ISCA Speech Synthesis Workshop (SSW), pp. 222–226 (2021). https://doi.org/10.21437/SSW.2021-39

  11. Ping, W., et al.: Deep voice 3: scaling text-to-speech with convolutional sequence learning. arXiv preprint arXiv:1710.07654 (2017)

  12. Ren, Y., et al.: Fastspeech: fast, robust and controllable text to speech. Adv. Neural Inf. Process. Syst. 32 (2019)

    Google Scholar 

  13. Shen, J., et al.: Natural TTS synthesis by conditioning wavenet on mel spectrogram predictions (2018). https://arxiv.org/abs/1712.05884

  14. Sun, M., Bellegarda, J.R.: Improved pos tagging for text-to-speech synthesis. In: 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 5384–5387. IEEE (2011)

    Google Scholar 

  15. Taylor, J., Richmond, K.: Analysis of pronunciation learning in end-to-end speech synthesis. In: INTERSPEECH, pp. 2070–2074 (2019)

    Google Scholar 

  16. Yao, K., Zweig, G.: Sequence-to-sequence neural net models for grapheme-to-phoneme conversion. arXiv preprint arXiv:1506.00196 (2015)

Download references

Acknowledgments

Supported by the ANR 19-P3IA-0003 MIAI. This work was performed using HPC/AI resources from GENCI-IDRIS (Grant AD011011542).

Author information

Authors and Affiliations

  1. Grenoble -Alps Univ, GIPSA-Lab, 11, rue des Mathématiques, St Martin d’Hères, France

    Maria-Loulou Hajj, Martin Lenglet, Olivier Perrotin & Gérard Bailly

Authors
  1. Maria-Loulou Hajj
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Martin Lenglet
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Olivier Perrotin
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Gérard Bailly
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Gérard Bailly .

Editor information

Editors and Affiliations

  1. Indian Institute of Technology Dharwad, Dharwad, India

    S. R. Mahadeva Prasanna

  2. St. Petersburg Federal Research Center of the Russian Academy of Sciences, St. Petersburg, Russia

    Alexey Karpov

  3. Koneru Lakshmaiah Education Foundation, Vaddeswaram, India

    K. Samudravijaya

  4. KIIT Group of Colleges, Gurugram, India

    Shyam S. Agrawal

A Appendices

A Appendices

1.1 A.1 Example of Embeddings of Word Pairs (B-wrd)

For example, processing two sentences using the word “as” by the LDA from FlauBERT embeddings of “as”:

figure m

1.2 A.2 Example of Embeddings of Class Pairs (B-grp)

Phonetization of heterophone homographes of a FaceBook post of French poetry with no errors:

figure n

Rights and permissions

Reprints and permissions

Copyright information

© 2022 Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Hajj, ML., Lenglet, M., Perrotin, O., Bailly, G. (2022). Comparing NLP Solutions for the Disambiguation of French Heterophonic Homographs for End-to-End TTS Systems. In: Prasanna, S.R.M., Karpov, A., Samudravijaya, K., Agrawal, S.S. (eds) Speech and Computer. SPECOM 2022. Lecture Notes in Computer Science(), vol 13721. Springer, Cham. https://doi.org/10.1007/978-3-031-20980-2_23

Download citation

  • DOI: https://doi.org/10.1007/978-3-031-20980-2_23

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-031-20979-6

  • Online ISBN: 978-3-031-20980-2

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics