Few-Shot Multimodal Learning

Meta-Learning

Approach where a model learns to learn by training on a variety of different tasks to develop rapid generalization capability to new tasks with few examples. Meta-learning optimizes model initialization parameters to facilitate rapid adaptation.

← Zurück