Intelligent Information Systems
-
Lecture 11 - Backpropagation Training
First page
Back
Continue
Last page
Overview
Graphics
Pattern vs Batch Training
Batch mode is closer to true gradient descent
requires smaller learning rate
Step size is smaller
Smooth traversal of the error surface
Requires many epochs
Tweet
Maintained by Dr Michael J. Watts (
mike.watts.net.nz
).