Using Fluency Representation Learned from Sequential Raw Features for Improving Non-native Fluency Scoring.
Kaiqi FuShaojun GaoXiaohai TianWei LiZejun MaPublished in: INTERSPEECH (2022)
Keyphrases
- representation scheme
- feature representation
- feature extraction
- raw data
- feature set
- low level
- search engine
- data sets
- probabilistic model
- co occurrence
- feature vectors
- image representation
- false positives
- classification method
- feature space
- extracted features
- computer vision
- temporal structure
- explicitly represented