Learning Speech Representations with Variational Predictive Coding
Transactions of the Association for Computational Linguistics · 01.07.2026 · 14 · 1489-1504
DOI: 10.1162/tacl.a.738 ↗Kurz gefasst
Fachartikel in der Zeitschrift Transactions of the Association for Computational Linguistics, online erschienen am 01.07.2026. Verfasst von Sung-Lin Yeh und 2 weiteren Autor:innen. Die Arbeit nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Despite being the best known objective for learning speech representations, the HuBERT objective has not been further developed and improved. We argue that it is the lack of an underlying principle that stalls the development, and, in this paper, we show that predictive coding under a variational view is the principle behind the HuBERT objective. Due to its generality, our formulation provides opportunities to improve parameterization and optimization, and we show two simple modifications that bring immediate improvements to the HuBERT objective. In addition, the predictive coding formulation has tight connections to various other objectives, such as APC, CPC, wav2vec, and BEST-RQ. Empirically, the improvement in pre-training brings significant improvements to four downstream tasks: phone classification, f0 tracking, speaker recognition, and automatic speech recognition, highlighting the importance of the predictive coding interpretation.
Zitieren
APA
Yeh, S., Bell, P., & Tang, H. (2026). Learning Speech Representations with Variational Predictive Coding. Transactions of the Association for Computational Linguistics, 14, 1489-1504. https://doi.org/10.1162/tacl.a.738
Vancouver
Yeh S, Bell P, Tang H. Learning Speech Representations with Variational Predictive Coding. Transactions of the Association for Computational Linguistics. 2026;14:1489-1504. doi:10.1162/tacl.a.738
BibTeX
@article{Yeh2026,
author = {Yeh, Sung-Lin and Bell, Peter and Tang, Hao},
title = {Learning Speech Representations with Variational Predictive Coding},
journal = {Transactions of the Association for Computational Linguistics},
year = {2026},
volume = {14},
pages = {1489--1504},
doi = {10.1162/tacl.a.738}
}
EUVIMED