← Back to glossary
Browse letter O hub
OOV (Out-of-Vocabulary)
Words or tokens that do not appear in a model’s training vocabulary and therefore cannot be directly recognised or translated by the system.
Definition
Words or tokens that do not appear in a model’s training vocabulary and therefore cannot be directly recognised or translated by the system.
How It Works
OOV (Out-of-Vocabulary) helps teams build predictable AI and translation workflows by setting clear expectations for quality, consistency, and decision-making.
In production environments, this concept is applied with process controls such as human review, terminology alignment, and repeatable quality checks across multilingual content.
Key Concepts
- core principle of oov (out-of-vocabulary)
- workflow-level implementation
- terminology and quality consistency
- human validation before publication
Where It Is Used
- localisation workflows
- AI translation pipelines
- multilingual content production
- cross-referencing related concepts such as OCR (Optical Character Recognition)