LibriSpeech-PC: Benchmark for Evaluation of Punctuation and Capitalization Capabilities of end-to-end ASR Models
This work addresses a domain-specific problem for ASR researchers by providing a benchmark to improve evaluation methods, though it is incremental as it builds on existing datasets and focuses on assessment rather than new model development.
The authors tackled the problem of evaluating punctuation and capitalization in end-to-end ASR models by introducing the LibriSpeech-PC benchmark, which includes a dataset with restored punctuation and capitalization, a novel Punctuation Error Rate (PER) metric, and baseline models, all made publicly available.
Traditional automatic speech recognition (ASR) models output lower-cased words without punctuation marks, which reduces readability and necessitates a subsequent text processing model to convert ASR transcripts into a proper format. Simultaneously, the development of end-to-end ASR models capable of predicting punctuation and capitalization presents several challenges, primarily due to limited data availability and shortcomings in the existing evaluation methods, such as inadequate assessment of punctuation prediction. In this paper, we introduce a LibriSpeech-PC benchmark designed to assess the punctuation and capitalization prediction capabilities of end-to-end ASR models. The benchmark includes a LibriSpeech-PC dataset with restored punctuation and capitalization, a novel evaluation metric called Punctuation Error Rate (PER) that focuses on punctuation marks, and initial baseline models. All code, data, and models are publicly available.