AI glossary
What is gradient descent?
Short answer
An optimisation method that repeatedly steps parameters opposite to the gradient to reduce loss.
For example
Training a model to predict house prices, gradient descent repeatedly adjusts each weight a little to shrink the gap between predicted and real prices.
Learn it in a free lesson
- Gradient descent: rolling downhillIntermediate · 6 min
- Backpropagation: blame, passed backwardsIntermediate · 6 min
- How machines learnBeginner · 5 min
Related terms
Questions
What is gradient descent in simple terms?
An optimisation method that repeatedly steps parameters opposite to the gradient to reduce loss.
What is an example of gradient descent?
Training a model to predict house prices, gradient descent repeatedly adjusts each weight a little to shrink the gap between predicted and real prices.
Where can I learn about gradient descent for free?
AI in Bed explains it in the free lesson "Gradient descent: rolling downhill" (Intermediate level, 6 minutes), with no sign-up.