Terra Incognita
Mathematics

Why do deep neural networks generalize so well?

Serious leads

Why it matters

Deep learning works very well in practice, but we do not yet understand why. Knowing why networks generalize could explain puzzling behaviours, such as learning from noisy labels, and suggest training methods that memorize less and generalize better.

What we know

Huge networks trained on finite data still work on new examples, contrary to classical statistics.

What blocks us

There is no complete theory explaining why, or predicting when it fails.

Current leads

Scaling laws, infinite-width limits, studies of learning dynamics.

Sub-questions

Related questions

Sources

Not yet reviewed by a specialist.

Last checked: 2026-11-18

History of this question

Back to the map →