SPGISpeech: 5,000 hours of transcribed financial audio for fully formatted end-to-end speech recognition
This work addresses the complexity and performance limitations in speech recognition for financial domains by integrating formatting directly into the model, though it is incremental as it builds on existing neural architectures.
The authors tackled the problem of speech-to-text transcription requiring separate post-processing for formatting by proposing an end-to-end neural model that outputs fully formatted text, achieving a character error rate of 1.7 on a 5,000-hour financial audio corpus.
In the English speech-to-text (STT) machine learning task, acoustic models are conventionally trained on uncased Latin characters, and any necessary orthography (such as capitalization, punctuation, and denormalization of non-standard words) is imputed by separate post-processing models. This adds complexity and limits performance, as many formatting tasks benefit from semantic information present in the acoustic signal but absent in transcription. Here we propose a new STT task: end-to-end neural transcription with fully formatted text for target labels. We present baseline Conformer-based models trained on a corpus of 5,000 hours of professionally transcribed earnings calls, achieving a CER of 1.7. As a contribution to the STT research community, we release the corpus free for non-commercial use at https://datasets.kensho.com/datasets/scribe.