Model Training & Adaptation
Bias-variance tradeoff
The tension between AI being too simple to learn useful patterns and too flexible to ignore accidental ones.
Example
A small AI system underfits while a complex one performs poorly on new cases.
Why people use it
It helps explain why both overly simple and overly complicated predictors can fail.
What you'll hear
“Are we missing the pattern or learning the noise?”
What this means for you
Choose complexity using performance on representative unseen data.
Can you control it?
No
No direct control. This describes a wider issue, concept or result rather than something you can simply switch on or off in a tool.
Common questions
- Is bias in this phrase the same as social unfairness?
- No. Here it refers to systematic modeling error, although fairness is a separate concern.
- Can a more complex system do worse on new examples?
- Yes. It can learn accidental details that do not carry over to new situations.
- Does one perfect training score settle the tradeoff?
- No. Performance on unfamiliar examples is needed to judge whether the learned pattern is useful.