Back to Glossary

Few-Shot Learning

NLP & Language Models

Learning from only a few examples.


Few-shot learning enables a model to solve new tasks with just a few examples. The model leverages its prior knowledge and generalization capabilities.

  • Examples: Classifying new categories, translating with limited examples.
  • Methods: In-context learning, fine-tuning, meta-learning.
  • Benefit: Efficient adaptation to small datasets.