VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretation

1 code implementation • ACL 2021 • Changhan Wang, Morgane Rivière, Ann Lee, Anne Wu, Chaitanya Talnikar, Daniel Haziza, Mary Williamson, Juan Pino, Emmanuel Dupoux

We introduce VoxPopuli, a large-scale multilingual corpus providing 100K hours of unlabelled speech data in 23 languages.

Representation Learning speech-recognition +1

Joint Masked CPC and CTC Training for ASR

1 code implementation • 30 Oct 2020 • Chaitanya Talnikar, Tatiana Likhomanenko, Ronan Collobert, Gabriel Synnaeve

Self-supervised learning (SSL) has shown promise in learning representations of audio that are useful for automatic speech recognition (ASR).

Automatic Speech Recognition Automatic Speech Recognition (ASR) +2

Linear Range in Gradient Descent

1 code implementation • 11 May 2019 • Angxiu Ni, Chaitanya Talnikar

This paper defines linear range as the range of parameter perturbations which lead to approximately linear perturbations in the states of a network.