Whispering LLaMA: A Cross-Modal Generative Error Correction Framework for Speech Recognition

Srijith Radhakrishnan, Chao-Han Yang, Sumeer Khan, Rohit Kumar, Narsis Kiani, David Gomez-Cabrero, Jesper Tegnér


Abstract
We introduce a new cross-modal fusion technique designed for generative error correction in automatic speech recognition (ASR). Our methodology leverages both acoustic information and external linguistic representations to generate accurate speech transcription contexts. This marks a step towards a fresh paradigm in generative error correction within the realm of n-best hypotheses. Unlike the existing ranking-based rescoring methods, our approach adeptly uses distinct initialization techniques and parameter-efficient algorithms to boost ASR performance derived from pre-trained speech and text models. Through evaluation across diverse ASR datasets, we assess our fusion technique, demonstrating a 37.66% improvement in word error rate (WER) relative performance compared to the n-best Oracle. To encourage future research, we have made our code and pre-trained models open source at [https://github.com/Srijith-rkr/Whispering-LLaMA](https://github.com/Srijith-rkr/Whispering-LLaMA)
Anthology ID:
2023.emnlp-main.618
Volume:
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
10007–10016
Language:
URL:
https://aclanthology.org/2023.emnlp-main.618
DOI:
10.18653/v1/2023.emnlp-main.618
Bibkey:
Cite (ACL):
Srijith Radhakrishnan, Chao-Han Yang, Sumeer Khan, Rohit Kumar, Narsis Kiani, David Gomez-Cabrero, and Jesper Tegnér. 2023. Whispering LLaMA: A Cross-Modal Generative Error Correction Framework for Speech Recognition. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 10007–10016, Singapore. Association for Computational Linguistics.
Cite (Informal):
Whispering LLaMA: A Cross-Modal Generative Error Correction Framework for Speech Recognition (Radhakrishnan et al., EMNLP 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.emnlp-main.618.pdf
Video:
 https://aclanthology.org/2023.emnlp-main.618.mp4