Glossary · AI & ML

Few-Shot Learning

An AI technique where a model learns to perform a task from just a few examples provided in the prompt.

Few-shot learning is a machine learning approach where a model learns to perform a task from a very small number of examples (typically 2-10), rather than requiring thousands of labeled training samples. In the context of large language models, few-shot learning involves providing examples in the prompt that demonstrate the desired input-output pattern. For text-to-SQL applications, few-shot examples might show natural language questions paired with their correct SQL translations for a specific database. This helps the model understand the schema context, naming conventions, and query patterns without additional fine-tuning.

Fig — every answer ships with the tables, rows and SQL behind it.