The SLT 2021 children speech recognition challenge: Open datasets, rules and baselines

TL;DR

This paper introduces the SLT 2021 Children Speech Recognition Challenge, providing open datasets, rules, and baselines to advance research in robust children's speech recognition using deep learning.

Contribution

It launches a large-scale Mandarin children speech dataset and a benchmarking challenge to improve CSR performance and robustness.

Findings

Release of 400 hours of Mandarin children's speech data

Establishment of challenge tracks and evaluation protocols

Provision of baseline models for CSR benchmarking

Abstract

Automatic speech recognition (ASR) has been significantly advanced with the use of deep learning and big data. However improving robustness, including achieving equally good performance on diverse speakers and accents, is still a challenging problem. In particular, the performance of children speech recognition (CSR) still lags behind due to 1) the speech and language characteristics of children's voice are substantially different from those of adults and 2) sizable open dataset for children speech is still not available in the research community. To address these problems, we launch the Children Speech Recognition Challenge (CSRC), as a flagship satellite event of IEEE SLT 2021 workshop. The challenge will release about 400 hours of Mandarin speech data for registered teams and set up two challenge tracks and provide a common testbed to benchmark the CSR performance. In this paper, we introduce the datasets, rules, evaluation method as well as baselines.