What is the gradient in Gradient Descent?

Updated May 16, 2026

Short answer

The gradient is a vector of partial derivatives indicating direction of steepest increase.

Deep explanation

The gradient shows how much the function changes with respect to each parameter. Gradient Descent uses the negative gradient to move toward minimum values.

Real-world example

Adjusting weights in neural networks based on error sensitivity.

Common mistakes

  • Ignoring that gradient is vector-valued in multivariate functions.

Follow-up questions

  • What is partial derivative?
  • What is gradient direction?

More Gradient Descent interview questions

View all →