Probabilistic asr feature extraction applying context-sensitive connectionist temporal classification networks.
Martin WöllmerBjörn W. SchullerGerhard RigollPublished in: ICASSP (2013)
Keyphrases
- context sensitive
- feature extraction
- feature vectors
- image classification
- feature space
- language model
- feature selection
- natural language
- graph grammars
- text classification
- machine learning
- decision trees
- connectionist networks
- context free
- multiple task learning
- model selection
- feature set
- information retrieval
- supervised learning
- speech recognition
- temporal information
- support vector machine
- extracted features
- spelling correction
- machine learning algorithms
- classification models
- temporal relations
- kernel principal component analysis
- active learning
- training set
- support vector