Prioritizing longest utterances in SSL speech pre-training data outperforms random or diversity-based sampling for ASR performance while using half the data volume.
Reducing barriers to self-supervised learning: Hubert pre-training with academic compute
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Study of Data Selection Strategies for Pre-training Self-Supervised Speech Models
Prioritizing longest utterances in SSL speech pre-training data outperforms random or diversity-based sampling for ASR performance while using half the data volume.