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 evaluate the stability and reproducibility of our fusion technique, demonstrating its improved word error rate relative (WERR) performance in comparison to n-best hypotheses by relatively 37.66%. To encourage future research, we have made our code and pre-trained models open source at https://github.com/Srijith-rkr/Whispering-LLaMA.
@article{arxiv.2310.06434,
title = {Whispering LLaMA: A Cross-Modal Generative Error Correction Framework for Speech Recognition},
author = {Srijith Radhakrishnan and Chao-Han Huck Yang and Sumeer Ahmad Khan and Rohit Kumar and Narsis A. Kiani and David Gomez-Cabrero and Jesper N. Tegner},
journal= {arXiv preprint arXiv:2310.06434},
year = {2025}
}
Comments
Accepted to EMNLP 2023 as main paper. 10 pages. Revised math notations. GitHub: https://github.com/Srijith-rkr/Whispering-LLaMA