Model Training & Adaptation
Data augmentation
Creating modified or additional training examples to improve robustness or coverage.
Example
An image dataset is expanded by rotating, cropping, and brightening existing pictures.
Why it matters
It exposes a system to useful variations without collecting every version separately.
You'll hear it
“Rotate the picture, but make sure the label still makes sense.”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Can a change accidentally alter the correct answer?
- Yes. Flipping, cropping or editing material can change its meaning or remove the relevant object.
- Does more augmentation always help?
- No. Unrealistic changes can teach patterns the system will not encounter in practice.
- Can it be used outside images?
- Yes. Suitable variations can also be created for sound, writing or other material.