Data & Databases
Labeled data
Examples with answers or descriptions attached, so AI can learn from them or be checked.
Example
People mark which objects appear in teaching photographs.
Why people use it
It gives AI examples with answers it can learn from or be tested against.
What you'll hear
“These photos already have checked labels.”
What this means for you
Check that people labeling the examples are following the same instructions.
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
- Are labels automatically correct?
- No. Labeling mistakes and inconsistent instructions can affect the collection of information.
- Can labels become outdated?
- Yes. Categories, rules or the basic facts can change after the labels were added.
- Must every label be supplied by a person?
- No. Labels can come from records, measurements or other systems, although their reliability still matters.