paper-with-me

Papers

Multi-head Monotonic Chunkwise Attention For Online Speech Recognition

2020-05-01 · Baiji Liu, Songjun Cao, Sining Sun, Weibin Zhang, Long Ma

The attention mechanism of the Listen, Attend and Spell (LAS) model requires the whole input sequence to calculate the attention context and thus is not suitable for online speech recognition. To deal with this problem, we propose multi-head monotonic chunk-wise attention (MTH-MoChA), an improved version of MoChA. MTH-MoChA splits the input sequence into small chunks and computes multi-head attentions over the chunks. We also explore useful training strategies such as LSTM pooling, minimum world error rate training and SpecAugment to further improve the performance of MTH-MoChA. Experiments on AISHELL-1 data show that the proposed model, along with the training strategies, improve the character error rate (CER) of MoChA from 8.96% to 7.68% on test set. On another 18000 hours in-car speech data set, MTH-MoChA obtains 7.28% CER, which is significantly better than a state-of-the-art hybrid system.

📄 PDF Abstract BibTeX arXiv:2005.00205

Code (0)

등록된 구현이 없습니다.

Tasks

speech-recognitionSpeech Recognition

Methods 이 논문이 사용한 방법론

Sigmoid Activation 설명 없음
Tanh Activation 설명 없음
LSTM An LSTM is a type of recurrent neural network that addresses the vanishing gradient problem in vanilla…

Similar Papers 제목 키워드 기반

Enhancing Monotonic Multihead Attention for Streaming ASR

2020-05-19 · Hirofumi Inaguma, Masato Mimura, Tatsuya Kawahara

We investigate a monotonic multihead attention (MMA) by extending hard monotonic attention to Transformer-based automatic speech recognition (ASR) for online streaming applications. For streaming inference, all monotonic…

AllAutomatic Speech RecognitionAutomatic Speech Recognition (ASR)Boundary Detection+2

Towards Online End-to-end Transformer Automatic Speech Recognition

2019-10-25 · Emiru Tsunoo, Yosuke Kashiwagi, Toshiyuki Kumakura, Shinji Watanabe

The Transformer self-attention network has recently shown promising performance as an alternative to recurrent neural networks in end-to-end (E2E) automatic speech recognition (ASR) systems. However, Transformer has a dr…

Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Decoderspeech-recognition+1

Monotonic Chunkwise Attention

2017-12-14 · Chung-Cheng Chiu, Colin Raffel

Sequence-to-sequence models with soft attention have been successfully applied to a wide variety of problems, but their decoding process incurs a quadratic time and space cost and is inapplicable to real-time sequence tr…

Document Summarizationspeech-recognitionSpeech Recognition

Monotonic Chunkwise Attention

2018-01-01 · ICLR 2018 1 · Chung-Cheng Chiu*, Colin Raffel*

Sequence-to-sequence models with soft attention have been successfully applied to a wide variety of problems, but their decoding process incurs a quadratic time and space cost and is inapplicable to real-time sequence tr…

Document Summarizationspeech-recognitionSpeech Recognition

Transformer-based Streaming ASR with Cumulative Attention

2022-03-11 · Mohan Li, Shucong Zhang, Catalin Zorila, Rama Doddipatla

In this paper, we propose an online attention mechanism, known as cumulative attention (CA), for streaming Transformer-based automatic speech recognition (ASR). Inspired by monotonic chunkwise attention (MoChA) and head-…

Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Decoderspeech-recognition+1