Answer: Gradient descent

Gradient descent is the optimization algorithm utilized to minimize a function by iteratively moving towards the direction of the steepest descent. L1 or Lasso and L2 or Ridge Regression serve as regularization techniques, while feature crosses are methods for generating synthetic features from two or more existing features. For more details, visit [https://builtin.com/data-science/gradient-descent](https://builtin.com/data-science/gradient-descent).

Author: LeetQuiz Editorial Team