Hanamaru 日本語にほんご × ML/AI
← All topics

Advanced Topics

Meta-Learning

メタラーニング

Meta-learning trains systems to adapt efficiently to new tasks. In few-shot learning, a small support set provides the examples needed for adaptation.

Japanese terms

  1. Meta-learning — メタラーニング: Learning methods that improve a system’s ability to learn new tasks from limited experience.
  2. Few-shot learning — 少数しょうすうショット学習がくしゅう: Learning or adapting from only a small number of labeled examples.
  3. Support set — サポートセット: The small labeled example set supplied for adaptation in a few-shot learning task.

Transfer learning reuses prior knowledge, while meta-learning explicitly optimizes the process of rapid adaptation.