Self-and-Mixed Attention Decoder with Deep Acoustic Structure for Transformer-Based LVCSR

• DOI: 10.21437/interspeech.2020-2556
• Publication Date: 2020-10-27T05:22:11Z
• Authors: Xinyuan Zhou; Grandee Lee; Emre Yılmaz; Yanhua Long; Jiaen Liang; Haizhou Li
• Keywords: FOS: Computer and information sciences; Sound (cs.SD); 03 medical and health sciences; Audio and Speech Processing (eess.AS); FOS: Electrical engineering, electronic engineering, information engineering; 0305 other medical science; Computer Science - Sound; Electrical Engineering and Systems Science - Audio and Speech Processing

Transformer has shown impressive performance in automatic speech recognition.It uses an encoder-decoder structure with self-attention to learn the relationship between high-level representation of source inputs and embedding target outputs.In this paper, we propose a novel decoder that features self-and-mixed attention (SMAD) deep acoustic (DAS) improve Transformer-based LVCSR.Specifically, introduce mechanism multi-layer for multiple levels abstraction.We also design mixed learns alignment different abstraction its corresponding linguistic information simultaneously shared space.The ASR experiments on Aishell-1 show proposed achieves CERs 4.8% dev set 5.1% test set, which are best reported results task our knowledge.