Few-Shot Learning

Learning from Few Examples

Few-shot classification.

N-Way K-Shot

N classes, K examples each. Test on new classes.

Matching Networks

Compare to support set. Attention-based.

Model-Agnostic

MAML: learn good initialization. Can adapt quickly.

Key Takeaways

1. Learn from few examples
2. N-way K-shot evaluation
3. MAML for fast adaptation