https://scholars.lib.ntu.edu.tw/handle/123456789/498646
標題: | Towards structured deep neural network for automatic speech recognition | 作者: | Liao, Y.-H. LIN-SHAN LEE HUNG-YI LEE |
關鍵字: | deep neural network; structured learning | 公開日期: | 2016 | 起(迄)頁: | 137-144 | 來源出版物: | 2015 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2015 - Proceedings | 摘要: | In this paper we propose the Structured Deep Neural Network (structured DNN) as a structured and deep learning framework. This approach can learn to find the best structured object (such as a label sequence) given a structured input (such as a vector sequence) by globally considering the mapping relationships between the structures rather than item by item. When automatic speech recognition is viewed as a special case of such a structured learning problem, where we have the acoustic vector sequence as the input and the phoneme label sequence as the output, it becomes possible to comprehensively learn utterance by utterance as a whole, rather than frame by frame. Structured Support Vector Machine (structured SVM) was proposed to perform ASR with structured learning previously, but limited by the linear nature of SVM. Here we propose structured DNN to use nonlinear transformations in multi-layers as a structured and deep learning approach. This approach was shown to beat structured SVM in preliminary experiments on TIMIT. © 2015 IEEE. |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-84964414244&doi=10.1109%2fASRU.2015.7404786&partnerID=40&md5=14af5601555218b97ab4fbe6c697baa6 | DOI: | 10.1109/ASRU.2015.7404786 |
顯示於: | 電機工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。