Data augmentation, feature combination, and multilingual neural networks to improve ASR and KWS performance for low-resource languages.
Zoltán TüskePavel GolikDavid NoldenRalf SchlüterHermann NeyPublished in: INTERSPEECH (2014)
Keyphrases
- data sets
- neural network
- data analysis
- databases
- data quality
- original data
- data collection
- pattern recognition
- image data
- data processing
- raw data
- database
- high quality
- data structure
- statistical analysis
- high dimensional data
- synthetic data
- resource allocation
- expressive power
- feature selection
- data mining
- spoken language