Paper Info
Title
FASTEMIT: LOW-LATENCY STREAMING ASR WITH SEQUENCE-LEVEL EMISSION REGULARIZATION
Abstract
Streaming automatic speech recognition (ASR) aims to emit each hypothesized word as quickly and accurately as possible. However, emitting fast without degrading quality, as measured by word error rate (WER), is highly challenging. Existing approaches including Early and Late Penalties [1] and Constrained Alignments [2, 3] penalize emission delay by manipulating per-token or per-frame probability prediction in sequence transducer models [4]. While being successful in reducing delay, these approaches suffer from significant accuracy regression and also require additional word alignment information from an existing model. In this work, we propose a sequence-level emission regularization method, named FastEmit, that applies latency regularization directly on per-sequence probability in training transducer models, and does not require any alignment. We demonstrate that FastEmit is more suitable to the sequence-level optimization of transducer models [4] for streaming ASR by applying it on various end-to-end streaming ASR networks including RNN-Transducer [5], Transformer-Transducer [6, 7], ConvNet-Transducer [8] and Conformer-Transducer [9]. We achieve 150 similar to 300ms latency reduction with significantly better accuracy over previous techniques on a Voice Search test set. FastEmit also improves streaming ASR accuracy from 4.4%/8.9% to 3.1%/7.5% WER, meanwhile reduces 90th percentile latency from 210ms to only 30ms on LibriSpeech.
Year
DOI
Venue
2021
10.1109/ICASSP39728.2021.9413803
2021 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP 2021)
DocType
Citations 
PageRank 
Conference
0
0.34
References 
Authors
0
11
Name
Order
Citations
PageRank
Jiahui Yu126025.83
Chung-Cheng Chiu224828.00
Bo Li320642.46
Shuo-Yiin Chang4274.71
Tara N. Sainath53497232.43
Yanzhang He66416.36
Arun Narayanan742532.99
Wei Han87513.10
Anmol Gulati901.01
Yonghui Wu10106572.78
Ruoming Pang11109292.99