Universum Autoencoder-Based Domain Adaptation for Speech Emotion Recognition.
Jun DengXinzhou XuZixing ZhangSascha FrühholzBjörn W. SchullerPublished in: IEEE Signal Process. Lett. (2017)
Keyphrases
- speech emotion recognition
- domain adaptation
- target domain
- semi supervised learning
- transfer learning
- cross domain
- labeled data
- semi supervised
- multiple sources
- unlabeled data
- training data
- domain specific
- test data
- training samples
- co training
- unsupervised learning
- machine learning
- text classification
- active learning
- supervised learning
- labeled training data
- learning problems
- learning tasks
- principal component analysis
- information extraction
- maximum likelihood
- sentiment classification
- labeled and unlabeled data
- model selection