Adaptive Sparse and Monotonic Attention for Transformer-based Automatic Speech Recognition.
Chendong ZhaoJianzong WangWenqi WeiXiaoyang QuHaoqian WangJing XiaoPublished in: DSAA (2022)
Keyphrases
- automatic speech recognition
- speech recognition
- speech signal
- recognition errors
- word error rate
- broadcast news
- hidden markov models
- conversational speech
- noisy environments
- speech retrieval
- spontaneous speech
- spoken words
- acoustic features
- word recognition
- speech corpus
- spoken document retrieval
- neural network
- focus of attention
- fault diagnosis
- sparse representation
- multi modal
- information retrieval