AI Glossary

Training Data

Training data is the collection of examples used to teach a machine-learning model how to recognize patterns or produce desired outputs.

Also known as: Training dataset; learning data

Why it matters in collision repair

An automotive model's performance depends on whether its training data accurately represents the vehicles, damage types, images, repair conditions, and markets where the model will be used.

Collision-industry example

A damage-detection model is trained using vehicle images that have been reviewed and labeled with the correct damaged components and damage types.