SESKStuff Everyone Should Know
SESK: Stuff Everyone Should Know · sesk.info/ai-overfitting-defined

A model scores great on its training data but badly on new data. What is that called?

Overfitting. It memorized the training data instead of generalizing.

Start playing free

  1. Underfitting
  2. Overfitting ✔
  3. Regularization

Why: Overfitting means the model matches the training data so closely, quirks and noise included, that it fails to make correct predictions on new data.

Huh, didn't know that: Overly complex models and unrepresentative training data are two common causes.

The dad joke
My model memorized every answer on the practice test. Then the real test had different questions. Classic overfitting.
Like this card?
Play the game

References

Answer: Overfitting. Overfitting means the model matches the training data so closely, quirks and noise included, that it fails to make correct predictions on new data.

  1. Google Machine Learning Glossary: overfitting developers.google.com/machine-learning/glossary
    Creating a model that matches the training data so closely that the model fails to make correct predictions on new data.
    Checked 2026-10-10.
  2. Google ML Crash Course: Overfitting developers.google.com/machine-learning/crash-course/overfitting/overfitting
    Overfitting can be detected by observing diverging loss curves for training and validation sets on a generalization curve.
    Checked 2026-10-10.

Think this answer is wrong? Tap "Challenge this answer" on the card in the game and tell us why.

Test yourself in the game

Artificial IntelligenceA model does poorly even on its own training data. What is the likely problem?Artificial IntelligenceTraining error is far lower than test error. Which problem does that point to?Artificial IntelligenceWhat does dropout randomly remove during a training step?

That's stuff everyone should know.