Data & Databases
Training data
The examples used to teach an AI how to behave or make predictions.
Example
A translation AI learns from collections of writing.
Why people use it
It supplies the examples an AI system learns from.
What you'll hear
“Where did its teaching examples come from?”
What this means for you
Record where teaching examples came from and keep a separate set for testing.
Can you control it?
Developer-only
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
- Is training data the same as a test set?
- No. Test data is held apart to assess performance on unseen examples.
- Does it include everything the AI later encounters?
- No. Later questions and new information can differ from the examples used during training.
- Can repeated examples affect learning?
- Yes. Repetition can give some patterns more influence than their importance warrants.