Multimodal and Multilingual Embeddings for Large-Scale Speech Mining

By Paul-Ambroise Duquenne, Hongyu Gong, Holger Schwenk. 2021

42 citations recorded by the source corpus.

Subfields: Multimodal Learning, Speech & Audio

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Multimodal Learning

1. Holger Schwenk
2. Yu Zhang
3. Alexis Conneau
4. Wei-Ning Hsu
5. Jiatao Gu
6. Mingxuan Wang

In Speech & Audio

1. Wei-Ning Hsu
2. Ye Jia
3. Yu Zhang
4. Alexis Conneau
5. Changhan Wang
6. Juan Pino
7. Tom Ko
8. Ankur Bapna

Citations in the corpus

• Textless Speech-to-Speech Translation on Real Data (2021, positive): “In addition, Duquenne et al. (2021) have demonstrated the first proof of concept of direct S2S mining without using ASR or MT systems.”
• CVSS Corpus and Massively Multilingual Speech-to-Speech Translation (2022, neutral): “Duquenne et al. (2021) showed a proof-of-concept of such an approach.”
• Leveraging Pseudo-labeled Data to Improve Direct Speech-to-Speech Translation (2022, methodology): “To tackle this challenge, [8] proposes an automatic data mining method to perform speech-to-speech mining.”
• SpeechMatrix: A Large-Scale Mined Corpus of Multilingual Speech-to-Speech Translations (2022, methodology): “data, we sample the training data from different languages with the same approach as (Duquenne et al., 2021).”
• Leveraging unsupervised and weakly-supervised data to improve direct speech-to-speech translation (2022, background): “[8, 29] demonstrated that it is possible to mine S2ST data from multilingual untranscribed speech corpora and improve the performance of the trained S2ST models.”