Simplification of MAML that performs a weight update towards the solution optimized on the current task. This simpler approach offers comparable performance with reduced computational complexity.

📖

termini

Cross-Domain Few-Shot Learning

Variant of few-shot learning where training and test tasks come from different domains. This setup evaluates the generalization capability of the meta-model beyond similar data distributions.

🔍

Glossario IA

Reptile Algorithm

Cross-Domain Few-Shot Learning

Nessun risultato trovato