Recurrent Neural Network Based Small-footprint Wake-up-word Speech Recognition System with a Score Calibration Method | |
Li, Chenxing1,2; Zhu, Lei3; Xu, Shuang1; Gao, Peng3; Xu, Bo1 | |
2018-08 | |
会议日期 | 2018-8 |
会议地点 | Beijing |
英文摘要 | In this paper, we propose a small-footprint wake-upword speech recognition (WUWSR) system based on long shortterm memory (LSTM) recurrent neural network, and we design a novel back-end calibration scoring method named modified zero normalization (MZN). First, LSTM is trained to predict posterior probability of context-dependent state. Next, MZN is adopted to transfer posterior probability to normalized score, which is then converted to confidence score by dynamic programming. Finally, a certain wake-up-word is recognized according to the confidence score. This WUWSR system can recognize multiple wake-up words and change wake-up words flexibly. This system can guarantee low latency by omitting decoding network. Equal error rate (EER) is adopted as the evaluation metric. Experimental results show that the proposed LSTM-based system achieves 33.33% relative improvement compared with a baseline system based on deep feed-forward neural network. Combining the front-end LSTM acoustic model with back-end MZN method, our WUWSR system can achieve 51.92% relative improvement. |
语种 | 英语 |
内容类型 | 会议论文 |
源URL | [http://ir.ia.ac.cn/handle/173211/39848] ![]() |
专题 | 数字内容技术与服务研究中心_智能技术与系统工程 |
通讯作者 | Li, Chenxing |
作者单位 | 1.Institute of Automation, Chinese Academy of Sciences 2.University of Chinese Academy of Sciences 3.AI Lab, Rokid Inc. |
推荐引用方式 GB/T 7714 | Li, Chenxing,Zhu, Lei,Xu, Shuang,et al. Recurrent Neural Network Based Small-footprint Wake-up-word Speech Recognition System with a Score Calibration Method[C]. 见:. Beijing. 2018-8. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论