Zero-Shot Classification
Definition
Zero-shot classification assigns data to categories without task-specific labeled training examples.
The assignment of inputs to classes for which the used model has received no task-specific labeled training examples, often via class descriptions.
Simply explained
You describe new categories to the model. It tries to assign suitable examples without having been trained on exactly this assignment before.
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
An operation assigns new ticket categories based on clear descriptions and verifies them against labeled cases.
Limits and pitfalls
- It is a concrete inference task, not identical to general zero-shot learning or a prompt without examples.
Origin & history
NLI-based language models made textual zero-shot classification practically accessible from 2019 onwards.
Use in mid-sized companies
- Possible application: An operation assigns new ticket categories based on clear descriptions and verifies them against labeled cases.
Frequently asked questions
What does zero-shot classification mean?
The assignment of inputs to classes for which the used model has received no task-specific labeled training examples, often via class descriptions. What should be considered? It is a concrete inference task, not identical to general zero-shot learning or a prompt without examples.
Sources
- Benchmarking Zero-shot Text Classification · Yin et al., Benchmarking Zero-shot Text Classification
Source checked on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results.
Editorially reviewed on: