gradient descent
/ˌɡreɪdiənt dɪˈsent/
An optimization algorithm that minimizes error iteratively
gradient descent in a sentence
“Gradient descent adjusts weights to reduce the loss function.”
Origin of gradient descent
From Latin gradus (step) + descendere (to go down), from de- (down) + scandere (to climb)
Related Words
overfitting
When a model learns noise instead of the underlying pattern
underfitting
When a model is too simple to capture the underlying pattern
hyperparameter
A parameter set before training begins, not learned from data
epoch
One complete pass through the entire training dataset
batch size
The number of samples processed before updating the model
loss function
A measure of how wrong the model's predictions are