On Pretraining Data Diversity for Self-Supervised Learning.
Hasan Abed Al Kader HammoudTuhin DasFabio PizzatiPhilip H. S. TorrAdel BibiBernard GhanemPublished in: CoRR (2024)
Keyphrases
- prior knowledge
- data sets
- background knowledge
- data quality
- data processing
- database
- data collection
- data analysis
- learning algorithm
- raw data
- high quality
- high dimensional data
- data points
- original data
- spatial data
- data distribution
- sensor data
- learning systems
- image data
- knowledge discovery
- end users
- active learning
- evolutionary algorithm
- data structure
- synthetic data
- data sources
- missing data
- data streams
- reinforcement learning
- training data
- noisy data
- neural network
- complex data
- learned models