Coarse-To-Fine And Cross-Lingual ASR Transfer

2021/09/02 by Peter Polák, Polák, Peter, Ondřej Bojar +1
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques #Speech Recognition and Synthesis #Speech and Audio Processing

paper · pdf · doi:10.48550/arxiv.2109.00916

openalex publication_date 2021/09/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

End-to-end neural automatic speech recognition systems achieved recently state-of-the-art results, but they require large datasets and extensive computing resources. Transfer learning has been proposed to overcome these difficulties even across languages, e.g., German ASR trained from an English model. We experiment with much less related languages, reusing an English model for Czech ASR. To simplify the transfer, we propose to use an intermediate alphabet, Czech without accents, and document that it is a highly effective strategy. The technique is also useful on Czech data alone, in the style of coarse-to-fine training. We achieve substantial eductions in training time as well as word error rate (WER).

Citations

• Improved Regularization of Convolutional Neural Networks with Cutout
• Stochastic Gradient Methods with Layer-wise Adaptive Moments for Training of Deep Networks

Related

• Cross-Language Transfer Learning, Continuous Learning, and Domain Adaptation for End-to-End Automatic Speech Recognition
• Cross-lingual Knowledge Transfer and Iterative Pseudo-labeling for Low-Resource Speech Recognition with Transducers
• Bootstrap an end-to-end ASR system by multilingual training, transfer\n learning, text-to-text mapping and synthetic audio
• A Survey of Multilingual Models for Automatic Speech Recognition
• Deciphering Speech: a Zero-Resource Approach to Cross-Lingual Transfer\n in ASR