We present a system for the Zero Resource Speech Challenge 2021, which combines a Contrastive Predictive Coding (CPC) with deep cluster. In deep cluster, we first prepare pseudo-labels obtained by clustering the outputs of a CPC network with k-means. Then, we train an additional autoregressive model to classify the previously obtained pseudo-labels in a supervised manner. Phoneme discriminative representation is achieved by executing the second-round clustering with the outputs of the final layer of the autoregressive model. We show that replacing a Transformer layer with a Conformer layer leads to a further gain in a lexical metric. Experimental results show that a relative improvement of 35% in a phonetic metric, 1.5% in the lexical metric, and 2.3% in a syntactic metric are achieved compared to a baseline method of CPC-small which is trained on LibriSpeech 460h data. We achieve top results in this challenge with the syntactic metric.
@article{arxiv.2107.05899,
title = {Speech Representation Learning Combining Conformer CPC with Deep Cluster for the ZeroSpeech Challenge 2021},
author = {Takashi Maekaku and Xuankai Chang and Yuya Fujita and Li-Wei Chen and Shinji Watanabe and Alexander Rudnicky},
journal= {arXiv preprint arXiv:2107.05899},
year = {2022}
}