Definition

Few-shot learning is a technique where a machine learning model is trained or prompted with a very small number of examples (shots) to perform a specific task. This allows the model to generalize to new, unseen data with limited prior exposure.

Why it matters (in Poovi’s context)

It’s a precursor and related concept to fine-tuning, demonstrating how providing a few examples in a prompt can significantly improve task performance. Fine-tuning can be seen as an extension of few-shot learning where examples are moved from the prompt to a training dataset.

Key properties or components

  • Limited Examples
  • Prompt-Based Learning
  • Generalization Ability

Contradictions or debates

None.

Sources