Reduce, Reuse, Recycle: Is Perturbed Data better than Other Language augmentation for Low Resource Self-Supervised Speech Models.
Asad UllahAlessandro RaganoAndrew HinesPublished in: CoRR (2023)
Keyphrases
- data sets
- experimental data
- data collection
- natural language
- original data
- synthetic data
- data sources
- prior knowledge
- statistical analysis
- data analysis
- data points
- knowledge discovery
- data processing
- database
- neural network
- information resources
- historical data
- data quality
- highly correlated
- language learning
- data warehouse
- high quality
- database systems