AIExplainer
Deep Learning Mathematics Intermediate 1 min read

What is Momentum?

A technique used in machine learning to help algorithms learn faster and more efficiently

Momentum is a method that adds a fraction of the previous update to the current update, helping the algorithm escape local minima and converge to the optimal solution

Imagine you're rolling a ball down a hill, and the ball is moving slowly at first, but as it gains momentum, it starts rolling faster and faster, helping it reach the bottom of the hill more quickly

Self-driving cars use momentum to improve the efficiency of their navigation systems, allowing them to learn from experience and adapt to new situations more quickly

Momentum is used in stochastic gradient descent and other optimization algorithms to improve the speed and accuracy of machine learning models

Some people think that momentum is only used in physics, but it's also a key concept in machine learning, where it helps algorithms learn and adapt more efficiently

The concept of momentum in machine learning was first introduced in the 1980s, and has since become a standard technique in many optimization algorithms

inertia velocity optimization technique

Three products for different needs — explore what’s relevant to you.