Back to Glossary
AI Technology

Zero-shot Learning

AI's ability to perform tasks it wasn't explicitly trained on.

Definition

Zero-shot learning refers to an AI model's ability to perform tasks or recognise categories it was not explicitly trained on. Large language models exhibit strong zero-shot capabilities, allowing them to extract information, classify documents, or answer questions about domains not in their training data. This contrasts with traditional machine learning that requires labelled examples for every task. Zero-shot abilities enable document intelligence systems to handle new document types without custom training.

See Zero-shot in action

Understanding the terminology is the first step. See how Conductor applies these concepts to solve real document intelligence challenges.

Request a demo