Learning Representation Models for Interpretable Automatic Phonemic Grouping of the Parkinsonian Speech
Published in Automatic Assessment of Parkinsonian Speech (AAPS 2024), Communications in Computer and Information Science, vol. 2646, 2026 · Conference paper
Recommended citation: Guerrero-López, A., Arias-Londoño, J. D., Shattuck-Hufnagel, S., & Godino-Llorente, J. I. (2026). Learning Representation Models for Interpretable Automatic Phonemic Grouping of the Parkinsonian Speech. In J. I. Godino Llorente (Ed.), Automatic Assessment of Parkinsonian Speech. AAPS 2024. Communications in Computer and Information Science, vol. 2646 (pp. 66–82). Springer, Cham. https://doi.org/10.1007/978-3-032-07083-8_5 https://doi.org/10.1007/978-3-032-07083-8_5
This proceedings paper presents DA-MARTA, a domain-adversarial Gaussian Mixture Variational Autoencoder for interpretable phonemic grouping across speech corpora and dialects. It is a separate contribution from the MARTA journal article, extending the research to domain adaptation and cross-dialect evaluation.
Springer published the chapter online on 20 November 2025 and assigns 2026 as the bibliographic year of the proceedings volume. The workshop took place on 21–22 November 2024.