A Representation Learning Perspective on the Importance of Train-Validation Splitting in Meta-Learning.
Nikunj SaunshiArushi GuptaWei HuPublished in: CoRR (2021)
Keyphrases
- meta learning
- learning tasks
- inductive learning
- meta knowledge
- learning algorithm
- reinforcement learning
- knowledge acquisition
- model selection
- machine learning algorithms
- learning process
- prior knowledge
- active learning
- unsupervised learning
- machine learning
- classification accuracy
- supervised learning
- feature set
- database systems