Vector-quantized neural networks for acoustic unit discovery in the ZeroSpeech 2020 challenge

In this paper, we explore vector quantization for acoustic unit discovery. Leveraging unlabelled data, we aim to learn discrete representations of speech that separate phonetic content from speaker-specific details. We propose two neural models to tackle this challenge - both use vector quantization to map continuous features to a finite set of codes. The first model is a type of vector-quantized variational autoencoder (VQ-VAE). The VQ-VAE encodes speech into a sequence of discrete units before reconstructing the audio waveform. Our second model combines vector quantization with contrastive predictive coding (VQ-CPC). The idea is to learn a representation of speech by predicting future acoustic units. We evaluate the models on English and Indonesian data for the ZeroSpeech 2020 challenge. In ABX phone discrimination tests, both models outperform all submissions to the 2019 and 2020 challenges, with a relative improvement of more than 30%. The models also perform competitively on a downstream voice conversion task. Of the two, VQ-CPC performs slightly better in general and is simpler and faster to train. Finally, probing experiments show that vector quantization is an effective bottleneck, forcing the models to discard speaker information.

[1]  Okko Johannes Räsänen,et al.  Computational modeling of phonetic and lexical learning in early language acquisition: Existing models and future directions , 2012, Speech Commun..

[2]  Karen Livescu,et al.  Unsupervised Pre-Training of Bidirectional Speech Encoders via Masked Reconstruction , 2020, ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[3]  Aren Jansen,et al.  The Zero Resource Speech Challenge 2015: Proposed Approaches and Results , 2016, SLTU.

[4]  Thomas Schatz,et al.  Neural network vs. HMM speech recognition systems as models of human cross-linguistic phonetic perception , 2018 .

[5]  Herman Kamper,et al.  Unsupervised Feature Learning for Speech Using Correspondence and Siamese Networks , 2020, IEEE Signal Processing Letters.

[6]  Gaetan Hadjeres,et al.  Vector Quantized Contrastive Predictive Coding for Template-based Music Generation , 2020, ArXiv.

[7]  Sakriani Sakti,et al.  The Zero Resource Speech Challenge 2019: TTS without T , 2019, INTERSPEECH.

[8]  Sanjeev Khudanpur,et al.  Unsupervised Learning of Acoustic Sub-word Units , 2008, ACL.

[9]  Alexander Kain,et al.  Spectral voice conversion for text-to-speech synthesis , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[10]  Haizhou Li,et al.  VQVAE Unsupervised Unit Discovery and Multi-scale Code2Spec Inverter for Zerospeech Challenge 2019 , 2019, INTERSPEECH.

[11]  Hao Tang,et al.  An Unsupervised Autoregressive Model for Speech Representation Learning , 2019, INTERSPEECH.

[12]  Oriol Vinyals,et al.  Neural Discrete Representation Learning , 2017, NIPS.

[13]  Thomas Hain,et al.  Unsupervised Acoustic Unit Representation Learning for Voice Conversion using WaveNet Auto-encoders , 2020, INTERSPEECH.

[14]  Lorenzo Rosasco,et al.  Discovering discrete subword units with binarized autoencoders and hidden-Markov-model encoders , 2015, INTERSPEECH.

[15]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[16]  Armand Joulin,et al.  Libri-Light: A Benchmark for ASR with Limited or No Supervision , 2020, ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[17]  James R. Glass,et al.  Unsupervised Lexicon Discovery from Acoustic Input , 2015, TACL.

[18]  Satoshi Nakamura,et al.  Learning Supervised Feature Transformations on Zero Resources for Improved Acoustic Unit Discovery , 2018, IEICE Trans. Inf. Syst..

[19]  Zhizheng Wu,et al.  Merlin: An Open Source Neural Network Speech Synthesis System , 2016, SSW.

[20]  S. Sakti,et al.  Development of HMM-based Indonesian Speech Synthesis , 2008 .

[21]  Aren Jansen,et al.  The zero resource speech challenge 2017 , 2017, 2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU).

[22]  Micha Elsner,et al.  Measuring the perceptual availability of phonological features during language acquisition using unsupervised binary stochastic autoencoders , 2019, NAACL.

[23]  Ron J. Weiss,et al.  Unsupervised Speech Representation Learning Using WaveNet Autoencoders , 2019, IEEE/ACM Transactions on Audio, Speech, and Language Processing.

[24]  Yoshua Bengio,et al.  Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation , 2013, ArXiv.

[25]  Nicolas Usunier,et al.  Joint Learning of Speaker and Phonetic Similarities with Siamese Networks , 2016, INTERSPEECH.

[26]  Thomas Drugman,et al.  Towards Achieving Robust Universal Neural Vocoding , 2018, INTERSPEECH.

[27]  James R. Glass,et al.  A Nonparametric Bayesian Approach to Acoustic Model Discovery , 2012, ACL.

[28]  Herbert Gish,et al.  Unsupervised training of an HMM-based self-organizing unit recognizer with applications to topic classification and keyword discovery , 2014, Comput. Speech Lang..

[29]  James Glass,et al.  Learning Hierarchical Discrete Linguistic Units from Visually-Grounded Speech , 2020, ICLR.

[30]  F. Pellegrino,et al.  Different languages, similar encoding efficiency: Comparable information rates across the human communicative niche , 2019, Science Advances.

[31]  Oriol Vinyals,et al.  Representation Learning with Contrastive Predictive Coding , 2018, ArXiv.

[32]  Lin-Shan Lee,et al.  Multi-target Voice Conversion without Parallel Data by Adversarially Learning Disentangled Audio Representations , 2018, INTERSPEECH.

[33]  Aren Jansen,et al.  Evaluating speech features with the minimal-pair ABX task: analysis of the classical MFC/PLP pipeline , 2013, INTERSPEECH.

[34]  Satoshi Nakamura,et al.  Development of Indonesian Large Vocabulary Continuous Speech Recognition System within A-STAR Project , 2008, IJCNLP.

[35]  Hao Wu,et al.  Mixed Precision Training , 2017, ICLR.

[36]  Alexei Baevski,et al.  vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations , 2019, ICLR.

[37]  Armand Joulin,et al.  Unsupervised Pretraining Transfers Well Across Languages , 2020, ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[38]  J. Flanagan.  Speech Analysis, Synthesis and Perception , 1971 .

[39]  Ewald van der Westhuizen,et al.  Unsupervised acoustic unit discovery for speech synthesis using discrete latent-variable neural networks , 2019, INTERSPEECH.

[40]  Heiga Zen,et al.  WaveNet: A Generative Model for Raw Audio , 2016, SSW.

[41]  Aurko Roy,et al.  Fast Decoding in Sequence Models using Discrete Latent Variables , 2018, ICML.

[42]  Adam Finkelstein,et al.  Fftnet: A Real-Time Speaker-Dependent Neural Vocoder , 2018, 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[43]  Lukás Burget,et al.  Variational Inference for Acoustic Unit Discovery , 2016, Workshop on Spoken Language Technologies for Under-resourced Languages.