Better May Not Be Fairer: Can Data Augmentation Mitigate Subgroup Degradation?
Ming-Chang ChiuPin-Yu ChenXuezhe MaPublished in: CoRR (2022)
Keyphrases
- data collection
- database
- raw data
- synthetic data
- data analysis
- data sets
- input data
- probability distribution
- high quality
- application domains
- original data
- case study
- feature selection
- experimental data
- computer systems
- knowledge discovery
- sensor data
- missing data
- statistical analysis
- data mining methods
- multi dimensional
- sensor networks
- image data
- data points
- end users
- relational databases
- video sequences
- data structure
- training data
- information systems
- machine learning
- real world
- neural network