Unformatted text preview: Lyla continues to use the standard performance function P = -½ y∗− y 2 to train the
neural net. The net she is training is shown on the next page and on the tear off sheet. Part A1 (6 points)
What is the new equation for δi = ∂P/∂z at an output node, where z is the weighted sum of the
inputs to the output node. Your equation should be in terms of y and y* and zero or more of the
input values and weights that appear in the diagram on the next page. 7 The neural net that Lyla is training looks like this: Part A2 (6 points)
What is the new equation for δA at internal node A? Your equation should be in terms of y, y*, and
zero or more of the input values and weights that appear in the diagram above. See the notes at
the end of the quiz if you need help wth the definition of δA 8 Part B (26 points)
Lyla runs one step...
View Full Document
This document was uploaded on 03/17/2014 for the course EECS 6.034 at MIT.
- Fall '10
- Artificial Intelligence