Learning layer weights on continuous features and reusing them for discrete token extraction, plus fine-tuning XLS-R with extra data, yields 44% relative CER reduction on ML-SUPERB and tops the challenge's single-system leaderboard.
Advances in Neural Information Processing Systems 33, 12449–12460 (2020)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Multilingual Speech Recognition Using Discrete Tokens with a Two-step Training Strategy
Learning layer weights on continuous features and reusing them for discrete token extraction, plus fine-tuning XLS-R with extra data, yields 44% relative CER reduction on ML-SUPERB and tops the challenge's single-system leaderboard.