Audio-Attention Discriminative Language Model for ASR Rescoring

• Published in: ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 7944 - 7948
• Main Authors: Gandhe, Ankur, Rastrow, Ariya
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-05-2020
• Subjects: Acoustics; Adaptation models; attention; Data models; end-to-end; Error analysis; Hidden Markov models; language modeling; Mathematical model; minimum word error rate; Training
• ISSN: 2379-190X
• DOI: 10.1109/ICASSP40776.2020.9054335

End-to-end approaches for automatic speech recognition (ASR) benefit from directly modeling the probability of the word sequence given the input audio stream in a single neural network. However, compared to conventional ASR systems, these models typically require more data to achieve comparable results. Well-known model adaptation techniques, to account for domain and style adaptation, are not easily applicable to end-to-end systems. Conventional HMM-based systems, on the other hand, have been optimized for various production environments and use cases. In this work, we propose to combine the benefits of end-to-end approaches with a conventional system using an attention-based discriminative language model that learns to rescore the output of a first-pass ASR system. We show that learning to rescore a list of potential ASR outputs is much simpler than learning to generate the hypothesis. The proposed model results in up to 8% improvement in word error rate even when the amount of training data is a fraction of data used for training the first-pass system.