Optimization

Explore the different variations of the gradient descent algorithm.

Optimization is the selection of the best element (with regard to some criterion) from a set of available alternatives.

In the simplest case, an optimization problem consists of maximizing or minimizing a real function by systematically choosing input values from within an allowed set and computing the value of the function.

In the case of machine learning, optimization refers to minimizing the loss function by systematically updating the network weights. Mathematically, this is expressed as:

w=argminwC(w)w^{\prime} = arg min_w C(w)

given a loss function CC and weights ww.

Intuitively, it can be thought of as descending a high-dimensional landscape. If we could project it in 2D plot, the height of the landscape would be the value of the loss function, and the horizontal axis would be the values of our weights w. Ultimately, the goal is to reach the bottom of the landscape by iteratively exploring the space around us.

Get hands-on with 1200+ tech skills courses.