Chapter 15, Artificial Intelligence Video Solutions, Invitation to Computer Science

Problem 1

a. Write a Winograd schema and the challenge question. Also give the correct answer. (See the Special Interest Box "Victory in the Turing Test" earlier in this chapter.)
b. Now change one word in your sentence from Exercise 1a so that the correct answer is different.

Adriano Chikande

Numerade Educator

01:56

Problem 2

Suppose that in a formal logic, $\operatorname{green}(x)$ means that $x$ has the attribute of being green, frog $(x)$ means that $x$ has the attribute of being a bullfrog, and $J$ stands for the specific entity Jeremiah. Translate the following formal statements into English:
a. $\operatorname{frog}(J)$
b. $(\forall x)(\operatorname{frog}(x) \rightarrow \operatorname{green}(x))$

Christopher Stanley

Numerade Educator

00:34

Problem 3

Draw a semantic net that incorporates the knowledge contained in the following paragraph:
If I had to describe what distinguishes a table from other pieces of fumiture, I guess I would say it has to have four legs and a flat top. The legs, of course, hold up the top. Nancy's table is made of maple, but mine is bigger and is wainut.

Gianna Calciano

Numerade Educator

00:40

Problem 4

a. Use an English-like formal language to represent the knowledge explicitly contained in the following semantic net:
b. Add to your list from Exercise 4a the knowledge that can be inferred from the semantic net.

Catelyn Jo

Numerade Educator

01:35

Problem 5

In the following neural network, which combinations of input values cause node $N 3$ to fire? Each input signal $\left(x_1, x_2, x_3, x_4\right)$ can have a value of either 0 or 1 .

Breanna Ollech

Numerade Educator

01:35

Problem 6

Assign weights and threshold values in the following neural network so that the output neuron fires only when $x_1$ and $x_3$ have the value 1 and $x_2$ has the value 0. Remember that weights can be negative.

Breanna Ollech

Numerade Educator

01:35

Problem 7

The truth table for the XOR operation is given in Figure 15.8. As noted in the text, a neural network for this operation requires a hidden layer of neurons. A partial solution to a neural network for $X O R$ is shown at the top of the next page. In this solution, all neurons have a threshold value of 1 and the inputs $x_1$ and $x_2$ are binary. Finish this network by finding weights for the connections and test that the resulting network implements the XOR operation. (Hint: All weights can be integer values.)

Breanna Ollech

Numerade Educator

05:01

Problem 8

Find some literature or product information on a mobile device or tablet $P C$ that allows pen-based handwritten entries. What sort of scheme does this system use for handwriting recognition? Does the system use a neural network? Does it require initial training on the user's handwriting?

Samuel Goyette

Numerade Educator

01:21

Problem 9

Ant colonies are an example from nature of swarm intelligence. Find two other examples of swarm intelligence seen in nature.

Rikhil Makwana

Numerade Educator

04:11

Problem 10

You are a knowledge engineer and have been assigned the task of developing a knowledge base for an expert system to advise on mortgage loan applications. What are some sample questions you would ask the loan manager at a bank?

Mihir Nayar

Numerade Educator

Problem 11

We described both forward chaining and backward chaining as techniques used by inference engines in rule-based systems. In Section 11.2.2, we described how a parser might analyze a programming statement to produce a parse tree. Does the method described in Chapter 11 correspond more closely to forward chaining or to backward chaining? Explain.

Check back soon!

01:34

Problem 12

A rule-based system for writing the screenplays for mystery movies contains the following assertions and rules:
The hero is a spy.
The heroine is an interpreter.
If the hero is a spy, then one scene should take place in Berlin and one in Paris.
If the heroine is an interpreter, then the heroine must speak English.
If the heroine is an interpreter, then the heroine must speak Russian.
If one scene should take place in Berlin, then there can be no car chase.
If there can be no car chase, then there can be no crash scene.
If one scene should take place in Berlin, then the hero is European.
If one scene should take place in Paris, then the hero must speak French.
Can the following assertion be inferred? Explain.
The hero must speak French and there can be no crash scene.

Ian Shi

Numerade Educator

00:12

Problem 13

In Exercise 12, is it possible to add the following assertion to the knowledge base? Why or why not?
The hero is American.

Ali Soave

Numerade Educator

Problem 14

If you studied Prolog in Chapter 10 and have a Prolog interpreter available, try implementing the rule-based system of Exercise 12 in Prolog and pose the queries of Exercises 12 and 13.

Check back soon!