AWhat changed
Trained a multilingual, multitask speech-recognition model on 680,000 hours of weakly supervised audio and reported improved zero-shot robustness across diverse test sets.
Scientific scope
Trained a multilingual, multitask speech-recognition model on 680,000 hours of weakly supervised audio and reported improved zero-shot robustness across diverse test sets. Scope limitation: Whisper's September 2022 release preceded the December 2022 arXiv posting. Reported robustness does not imply the best specialized model on every speech benchmark.
Evidence
The retained primary-source abstract, bibliographic record and stated review scope support this provisional assessment. No independent experimental replication by WRS is claimed. Whisper's September 2022 release preceded the December 2022 arXiv posting. Reported robustness does not imply the best specialized model on every speech benchmark.
Contributors
- The named authors listed in the linked research record · SCIENCE: Complete source byline retained in the research record; human research direction, method development and evaluation. Individual labor shares are not measured.
- Radford and colleagues: Whisper robust speech recognition — original paper
- Radford and colleagues: Whisper robust speech recognition — company research page
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction