Meta-Learning for Few-Shot Classification

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.

← Kembali