Saved in:
Bibliographic Details
Main Authors: Kong, Xiangzhu, Ning, Tianqi, Huang, Hao, Ou, Zhijian
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.09807
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913503010357248
author Kong, Xiangzhu
Ning, Tianqi
Huang, Hao
Ou, Zhijian
author_facet Kong, Xiangzhu
Ning, Tianqi
Huang, Hao
Ou, Zhijian
contents Recently multi-channel end-to-end (ME2E) ASR systems have emerged. While streaming single-channel end-to-end ASR has been extensively studied, streaming ME2E ASR is limited in exploration. Additionally, recent studies call attention to the gap between in-distribution (ID) and out-of-distribution (OOD) tests and doing realistic evaluations. This paper focuses on two research problems: realizing streaming ME2E ASR and improving OOD generalization. We propose the CUSIDE-array method, which integrates the recent CUSIDE methodology (Chunking, Simulating Future Context and Decoding) into the neural beamformer approach of ME2E ASR. It enables streaming processing of both front-end and back-end with a total latency of 402ms. The CUSIDE-array ME2E models are shown to achieve superior streaming results in both ID and OOD tests. Realistic evaluations confirm the advantage of CUSIDE-array in its capability to consume single-channel data to improve OOD generalization via back-end pre-training and ME2E fine-tuning.
format Preprint
id arxiv_https___arxiv_org_abs_2407_09807
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CUSIDE-array: A Streaming Multi-Channel End-to-End Speech Recognition System with Realistic Evaluations
Kong, Xiangzhu
Ning, Tianqi
Huang, Hao
Ou, Zhijian
Audio and Speech Processing
Recently multi-channel end-to-end (ME2E) ASR systems have emerged. While streaming single-channel end-to-end ASR has been extensively studied, streaming ME2E ASR is limited in exploration. Additionally, recent studies call attention to the gap between in-distribution (ID) and out-of-distribution (OOD) tests and doing realistic evaluations. This paper focuses on two research problems: realizing streaming ME2E ASR and improving OOD generalization. We propose the CUSIDE-array method, which integrates the recent CUSIDE methodology (Chunking, Simulating Future Context and Decoding) into the neural beamformer approach of ME2E ASR. It enables streaming processing of both front-end and back-end with a total latency of 402ms. The CUSIDE-array ME2E models are shown to achieve superior streaming results in both ID and OOD tests. Realistic evaluations confirm the advantage of CUSIDE-array in its capability to consume single-channel data to improve OOD generalization via back-end pre-training and ME2E fine-tuning.
title CUSIDE-array: A Streaming Multi-Channel End-to-End Speech Recognition System with Realistic Evaluations
topic Audio and Speech Processing
url https://arxiv.org/abs/2407.09807