tut5 - 2 = 1 Compare it with the actual network outputs...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
EE4210 Tutorial 5 1. Use Generalized Delta Rule in Backpropagation Training (a) Complete the error backpropagation phase in Q2 of Tutorial 4 and update all the weights and thresholds in the multilayer perceptron. But this time, use the generalized delta rule with momentum constant α = 0.6. (b) Then use the new weights and thresholds to find the new actual network output y Y 1 for the same inputs X 1 = 0 and X
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 2 = 1. Compare it with the actual network outputs found in Q1 and Q2 of Tutorial 4. 2. Delta-Delta Rule for Learning Rate Adaptation Based on the trainings in Q1 of Tutorial 4 and Q1 of this tutorial, calculate the new learning rate parameter η for each weight/threshold if the delta-delta rule is employed after the second iteration. The control step-size parameter γ is chosen as 0.8....
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online