Combining Adversarial Training And Disentangled Speech Representation For Robust Zero-resource Subword Modeling
2019 · Siyuan Feng, Tan Lee, Zhiyuan Peng
Abstract
This study addresses the problem of unsupervised subword unit discovery from untranscribed speech. It forms the basis of the ultimate goal of ZeroSpeech 2019, building text-to-speech systems without text labels. In this work, unit discovery is formulated as a pipeline of phonetically discriminative feature learning and unit inference. One major difficulty in robust unsupervised feature learning is dealing with speaker variation. Here the robustness towards speaker variation is achieved by applying adversarial training and FHVAE based disentangled speech representation learning. A comparison of the two approaches as well as their combination is studied in a DNN-bottleneck feature (DNN-BNF) architecture. Experiments are conducted on ZeroSpeech 2019 and 2017. Experimental results on ZeroSpeech 2017 show that both approaches are effective while the latter is more prominent, and that their combination brings further marginal improvement in across-speaker condition. Results on ZeroSpeech 201
Authors
(none)
Tags
Stats
Related papers
- Improving Unsupervised Subword Modeling Via Disentangled Speech Representation Learning And Transformation (2019)5.24
- The Zero Resource Speech Challenge 2020: Discovering Discrete Subword And Word Units (2020)11.58
- Unsupervised Acoustic Unit Discovery For Speech Synthesis Using Discrete Latent-variable Neural Networks (2019)9.59
- Robust Disentangled Variational Speech Representation Learning For Zero-shot Voice Conversion (2022)10.97
- Multilingual And Unsupervised Subword Modeling For Zero-resource Languages (2018)7.81
- Transformer VQ-VAE For Unsupervised Unit Discovery And Speech Synthesis: Zerospeech 2020 Challenge (2020)9.41
- Unsupervised Neural And Bayesian Models For Zero-resource Speech Processing (2017)0.00
- Exploiting Cross-lingual Speaker And Phonetic Diversity For Unsupervised Subword Modeling (2019)6.77