← Back to Glossary
Learning
Zero-shot Learning
Completing new tasks without prior examples.
Zero-shot Learning refers to an AI model's ability to perform a task it has never explicitly seen during its training. Thanks to its vast general knowledge, it can deduce how to solve new problems based on the instructions given at the moment.
Examples
- Asking a model to classify tweets as 'positive/negative' without ever teaching it tweet examples.
- Translating between two rare languages the model never saw paired, using a bridge language like English.
Use Cases
- Real-time data classification without re-training.
- Rapid model adaptation to new business domains.
Next paths
Go deeper from this topic
Move from educational content into the product, solution, and implementation pages that matter next.