We introduce Generative Spoken Language Modeling, the task of learning the acoustic and linguistic characteristics of a language from raw audio (no text, no labels), and a set of metrics to automatically evaluate the learned representations at acoustic and linguistic levels for both encoding and generation. We set up baseline systems consisting of a discrete speech encoder (returning pseudo-text units), a generative language model (trained on pseudo-text), and a speech decoder (generating a waveform from pseudo-text) all trained without supervision and validate the proposed metrics with human evaluation. Across 3 speech encoders (CPC, wav2vec 2.0, HuBERT), we find that the number of discrete units (50, 100, or 200) matters in a task-dependent and encoder-dependent way, and that some combinations approach text-based systems.
@article{arxiv.2102.01192,
title = {Generative Spoken Language Modeling from Raw Audio},
author = {Kushal Lakhotia and Evgeny Kharitonov and Wei-Ning Hsu and Yossi Adi and Adam Polyak and Benjamin Bolte and Tu-Anh Nguyen and Jade Copet and Alexei Baevski and Adelrahman Mohamed and Emmanuel Dupoux},
journal= {arXiv preprint arXiv:2102.01192},
year = {2021}
}