Elastic Weight Consolidation Improves the Robustness of Self-Supervised Learning Methods under Transfer.
Andrius OvsianasJason RamapuramDan BusbridgeEeshan Gunesh DhekaneRuss WebbPublished in: CoRR (2022)
Keyphrases
- learning algorithm
- preprocessing
- empirical studies
- learning models
- human experts
- data sets
- learning process
- prior knowledge
- computational cost
- unsupervised learning
- learning systems
- statistical methods
- learning problems
- knowledge transfer
- inductive learning
- knowledge acquisition
- significant improvement
- support vector
- knowledge base