Beyond Simple Meta-Learning: Multi-Purpose Models for Multi-Domain, Active and Continual Few-Shot Learning.
Peyman BateniJarred BarberRaghav GoyalVaden MasraniJan-Willem van de MeentLeonid SigalFrank WoodPublished in: CoRR (2022)
Keyphrases
- meta learning
- learning tasks
- multi domain
- inductive learning
- model selection
- meta knowledge
- prior knowledge
- machine learning algorithms
- supervised learning
- learning algorithm
- learning process
- data sets
- unsupervised learning
- feature selection
- multi class
- domain specific
- classification accuracy
- active learning
- reinforcement learning
- case study