Zero-Shot Learning
Definition
Zero-shot learning recognizes previously unseen classes using additional descriptions or attributes.
A learning setting in which a model recognizes instances of previously unseen classes using side information such as attributes or semantic descriptions.
Simply explained
The model does not yet know a new category from examples. A description of its features still helps it recognize it.
Why it matters in practice
The term helps to select, develop, evaluate AI systems in a well-founded manner and realistically assess their technical risks.
Example from business
A spare parts model recognizes a new class based on defined properties and is subsequently validated in reality.
Limits and pitfalls
- Unseen classes require a semantic bridge; generalized ZSL and pure zero-shot classification must be evaluated differently.
Origin & history
Lampert and co-authors demonstrated attribute-based zero-shot learning for object classes in 2009.
Use in mid-sized companies
- Possible application: A spare parts model recognizes a new class based on defined properties and is subsequently validated in reality.
Frequently asked questions
What does learning without task-specific examples (zero-shot learning) mean?
A learning setting in which a model recognizes instances of previously unseen classes using side information such as attributes or semantic descriptions. What should be considered? Unseen classes require a semantic bridge; generalized ZSL and pure zero-shot classification must be evaluated differently.
Sources
- Learning to Detect Unseen Object Classes by Between-Class Attribute Transfer · Lampert et al., Learning to Detect Unseen Object Classes by Between-Class Attribute Transfer
Source verified on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results.
Editorially reviewed on: