What is gradient descent in cost optimization?

Updated May 15, 2026

Short answer

Gradient descent iteratively reduces cost by updating parameters in the opposite direction of the gradient.

Deep explanation

It computes partial derivatives of the cost function and updates parameters step by step to reach a minimum.

Real-world example

Used in training neural networks for image recognition.

Common mistakes

  • Using too high learning rate causing divergence.

Follow-up questions

  • What is learning rate?
  • What are local minima?

More Cost Function interview questions

View all →