A bilevel training method that jointly optimizes supervised and unsupervised losses outperforms pretraining-then-finetuning for ASR on LibriSpeech, Switchboard, and an in-house dataset.
Re- alizing petabyte scale acoustic modeling,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Bilevel Joint Unsupervised and Supervised Training for Automatic Speech Recognition
A bilevel training method that jointly optimizes supervised and unsupervised losses outperforms pretraining-then-finetuning for ASR on LibriSpeech, Switchboard, and an in-house dataset.