Model Training & Adaptation
Boosting
Building a series of prediction systems, with later ones helping correct earlier mistakes.
Example
A series of weak prediction systems is combined into a stronger predictor.
Why people use it
It combines simpler predictors while concentrating learning on earlier weaknesses.
What you'll hear
“The next part works on the earlier mistakes.”
What this means for you
Compare results on unfamiliar examples, rather than only how well the system learned its practice material.
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
- Do all the prediction systems learn independently?
- No. Later ones respond to earlier mistakes, so their learning is connected.
- Can it focus too much on bad examples?
- Yes. Incorrect or unusual examples may attract attention and make the final system less useful.
- Is it always better than one strong predictor?
- No. The answer depends on the task and the quality of the teaching examples.