Locality enhanced dynamic biasing and sampling strategies for contextual ASR
Abstract
Automatic Speech Recognition (ASR) still face challenges when recognizing time-variant rare-phrases. Contextual biasing (CB) modules bias ASR model towards such contextually-relevant phrases. During training, a list of biasing phrases are selected from a large pool of phrases following a sampling strategy. In this work we firstly analyse different sampling strategies to provide insights into the training of CB for ASR with correlation plots between the bias embeddings among various training stages. Secondly, we introduce a neighbourhood attention (NA) that localizes self attention (SA) to the nearest neighbouring frames to further refine the CB output. The results show that this proposed approach provides on average a 25.84% relative WER improvement on LibriSpeech sets and rare-word evaluation compared to the baseline.
Cite
@article{arxiv.2401.13146,
title = {Locality enhanced dynamic biasing and sampling strategies for contextual ASR},
author = {Md Asif Jalal and Pablo Peso Parada and George Pavlidis and Vasileios Moschopoulos and Karthikeyan Saravanan and Chrysovalantis-Giorgos Kontoulis and Jisi Zhang and Anastasios Drosou and Gil Ho Lee and Jungin Lee and Seokyeong Jung},
journal= {arXiv preprint arXiv:2401.13146},
year = {2024}
}
Comments
Accepted for IEEE ASRU 2023