Dropout
A training trick that randomly switches off a fraction of a network's neurons on each step, forcing it not to lean on any single one. It's one of the simplest ways to fight overfitting.
Think of It Like This
Cross-training a team so no single person is a point of failure.
On every training pass, dropout zeroes out a random subset of activations — say 20% of them — so the network has to spread what it learns across many neurons instead of a critical few. That forced redundancy is what makes it generalize better and overfit less. At inference time dropout is switched off and the full network runs, giving a smooth, ensemble-like average of all those thinned sub-networks.