This preview shows page 1. Sign up to view the full content.
Unformatted text preview: (18 Points)
Run back-propagation using these same parameter values for one iteration to determine new weight
values for WAD , WBD , and WCD . Use your results from Part A and Part B, and assume a learning rate of
r = 2 and a desired output, d=1. Part D: (8 Points)
Hermione changes all the decision functions in the original network, before training, from using a
sigmoid function to using a threshold function (which outputs 1 if the input is greater than 0, and
outputs 0 otherwise).
Using X, Y, and Z and the standard logical operators (AND, NOT, and OR), determine a logic
expression for the neural net. Assume that the inputs are always 0 (meaning false) or 1 (meaning true). 23 Quiz 4, Problem 1: SVMs (50 points) Part A: F.R.O.S.T. and the Snowmen (25 Points)
You have been hired by the Foundation for Research into Occult Snow Transformations (F.R.O.S.T.) to
investigate a strange new phenomenon. It seems that certain snowmen can be turned into living, talking
creatures when an old silk hat is placed on their heads. So far, F.R.O.S.T. researchers have determined
that the factors that b...
View Full Document
This document was uploaded on 03/17/2014 for the course EECS 6.034 at MIT.
- Fall '10
- Artificial Intelligence