Earnings22
A benchmark dataset designed for evaluating speech recognition in the financial domain, composed of real earnings call recordings from publicly listed companies. It contains abundant industry terminology, numerical expressions, and non-native accents, serving as a standard test set for measuring ASR system generalization in professional speech scenarios.
Timeline (last 90 days)
When adapting the general-domain trained Whisper-small model to the Earnings22 benchmark dataset, all evaluation methods showed improvement rather than degradation on the in-domain held-out set of financial speech during the adaptation process, and the expected catastrophic forgetting did not occur.
Earnings22 is a benchmark dataset specifically designed for speech recognition evaluation in the financial domain, consisting of real earnings call recordings from listed companies, containing a large number of industry terms, numerical expressions, and non-native accents, with word error rates generally higher than in general-domain scenarios.
All Facts (2)
When adapting the general-domain trained Whisper-small model to the Earnings22 benchmark dataset, all evaluation methods showed improvement rather than degradation on the in-domain held-out set of financial speech during the adaptation process, and the expected catastrophic forgetting did not occur.
50%UnverifiedEarnings22 is a benchmark dataset specifically designed for speech recognition evaluation in the financial domain, consisting of real earnings call recordings from listed companies, containing a large number of industry terms, numerical expressions, and non-native accents, with word error rates generally higher than in general-domain scenarios.
50%