A Representation Learning Perspective on the Importance of Train-Validation Splitting in Meta-Learning.
Nikunj SaunshiArushi GuptaWei HuPublished in: ICML (2021)
Keyphrases
- meta learning
- learning tasks
- inductive learning
- meta knowledge
- learning process
- learning algorithm
- feature selection
- active learning
- data sets
- supervised learning
- unsupervised learning
- semi supervised
- metamodel
- machine learning algorithms
- model selection
- software engineering
- expert systems
- reinforcement learning
- decision trees
- knowledge base
- databases