Regularising for invariance to data augmentation improves supervised learning.
Aleksander BotevMatthias BauerSoham DePublished in: CoRR (2022)
Keyphrases
- supervised learning
- data sets
- high quality
- training data
- data quality
- synthetic data
- prior knowledge
- data points
- spatial data
- data sources
- application domains
- database
- data collection
- sensor data
- experimental data
- machine learning
- knowledge discovery
- data analysis
- raw data
- data acquisition
- statistical analysis
- missing data
- training samples
- labeled data
- image data
- data structure
- neural network