A Quick Guide to Gradient Descent and its Variants

Author(s): Riccardo Di Sipio

Originally published on Towards AI.

And if you do not know where to start from, optimize the learning rate

In machine learning, one deals with some mathematical model z = F(x,θ) which is a function of some input variables x and a set of parameters θ. The model can be for example an artificial and possibly deep neural network. The name of the game is to find the set of parameters that minimize the error in predicting some targets y (or labels in classification problems) starting from some defined input x, i.e. to reduce the quantity F(x,θ)-y as much as possible. Borrowing the jargon and some ideas from optimization theory, if the model F(x,θ) is a differentiable function, i.e…. Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI