thinQmaster Virtual testing partner Adaptive practice exams that find your weak spots and drill them. Every answer explained. From $69
Enroll

thinQmaster

Find your exam

Five questions free, before you decide.

Step 1 · choose a vendor

In one of our camps? Your practice exams come with the tuition — they are on your study page already.

AI & Machine Learning · Intermediate

Stochastic Gradient Descent (SGD)

Gradient descent that updates parameters using one example or a small mini-batch at a time.

Definition

SGD makes frequent, noisy updates that are computationally cheap and can help escape shallow local minima. Mini-batch SGD, which uses batches of tens to thousands of examples, is the standard practical form. Momentum is often added to smooth updates.

Exam tip

Mini-batch is the practical compromise between full-batch stability and single-example speed.

Related terms