What is Gradient Descent?
Updated May 17, 2026
Short answer
Gradient descent minimizes the loss function iteratively.
Deep explanation
It updates coefficients in the direction of steepest loss reduction.
Real-world example
Large-scale recommendation systems use stochastic gradient descent.
Common mistakes
- Using inappropriate learning rates.
Follow-up questions
- What is SGD?
- What causes divergence?