Unsupervised Learning of Efficient and Robust Speech Representations

By Kazuya Kawakami, Luyu Wang, Chris Dyer, Phil Blunsom, Aäron van den Oord. 2019

7 citations recorded by the source corpus.

Subfields: Representation & Self-Supervised 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 Representation & Self-Supervised Learning

1. Alexei Baevski
2. Michael Auli
3. Abdel-rahman Mohamed
4. Hung-yi Lee
5. Andy T. Liu
6. Po-Han Chi
7. Po-Chun Hsu
8. Shu-Wen Yang

In Speech & Audio

1. Hung-yi Lee
2. Abdel-rahman Mohamed
3. Alexei Baevski
4. Michael Auli
5. Yu-An Chung
6. Andy T. Liu
7. Wei-Ning Hsu
8. James R. Glass

Citations in the corpus

• Effectiveness of self-supervised pre-training for speech recognition (2019, methodology): “The closest setup to ours is [38] which learns representations using CPC [11] and then feed these into an ASR system trained on about 96h of labeled data, which is surpassed on test-clean by our Discrete-BERT model trained on vq-wav2vec representations fine-tuned on 10mins and 1 hour.”
• Mockingjay: Unsupervised Speech Representation Learning with Deep Bidirectional Transformer Encoders (2019, background): “Unsupervised speech representation learning [2, 3, 4, 5, 6, 7, 8, 9, 10] is effective in extracting high-level properties from speech.”
• Effectiveness of Self-Supervised Pre-Training for ASR (2020, methodology): “The closest setup to ours is [39] which learns representations using CPC [11] and then feed these into an ASR system trained on about 96h of labeled data achieving 9.”