Lab 8 Artificial Intelligence

CS1150 Introduction to Computer Science

Lab #8 – Artificial Intelligence, 40 points

Objectives:

1. To learn how semantic networks and rule based natural language systems can simulate intelligent behavior.

2. To experiment with Chatbots to determine their authenticity.

3. To understand how to evaluate web resources to be able to discern the legitimacy of the source.

Instructions:

Make sure to do the Activity as outlined in the LabCH17.pdf file. The activity is actually a tutorial that will help you to solve the problems in the exercises. It is expected that you will complete the Activity before you begin the exercises.

Record your answers to the questions in a separate Word document with each question labelled to match this document. When you are finished, save and the completed document as “firstname_lastname_lab8.docx”. Then, submit the saved document to Pilot.

EXERCISE 1

Part 1

1. Start the “Semantic networks” applet.

2. Hit “Examples” to fill in some rules to start you off.

3. Add a new “isa” rule to the rule base. This rule may extend the human/animal categories or do something entirely different. [Example: An animal is a pizza]

Type your new rule here, below this line:

4. Add a new rule that uses a verb other than “is” and mentions your new category. [Example: a dog jumps]

Type your new rule here, below this line:

5. Add a new rule that gives a characteristic of your new category. This rule puts your new category in front, in the form of a possessive noun. [Example: dog has four legs]

Type your new rule here, below this line:

6. Type in one query that makes use of your rules, and which should evaluate to “true.” Take a screenshot showing the query, all the rules you made, and the answer the applet gave you.

Copy the screenshot that shows all those things here, below this line:

7. Type in one query that makes use of your rules, and which should evaluate to “false.” Take a screenshot showing the query, all the rules you made, and the answer the applet gave you.

Copy the screenshot that shows all those things here, below this line:

EXERCISE 1

Part 2

1. Start the Eliza Therapist app.

2. Click on Show rules.

3. Type in the rule that changes the client ‘s statement (Hint: Use variables)

My _______ is ruined

into this statement:

Are you sure your _______ is ruined

4. Test out your new rule pretending to be client.

5. Add a new rule that changes the client’s statement

I have a _____ for a pet

into this statement:

It must be interesting to have a ________

The first blank, however, cannot be the word “cat.” (Hint: Use variables and write a rule when the word cat is used, you won’t get the given response)

6. Test this new rule. Take a screenshot showing the rules and the above two

statements, along with Eliza’s responses.

Put your screenshot below this line:

7. Type in a statement that makes Eliza come up with a contextually incorrect

Response. Take a screenshot. (e.g. – If you ask Eliza “How are you” and she

responds as “Weather is nice”)

Put your screenshot below this line:

Exercise 2

Working with Chatbots

What is a Chatbot?

A chatbot is a computer program that tries to engage in a conversation with the user who interacts with it. The conversations between a chatbot and a user can take place using textual inputs via a keyboard or audio inputs via a microphone. The task of the chatbot is to mimic humans when answering the questions in a way that the user who interacts with the chatbot will not be able find out that he/she is actually conversing with a computer program. Today, chatbots or conversational agents are a part of many real-world applications such as mobile phone assistants (e.g., Apple’s Siri, Microsoft’s Cortana, and Google Assistant), virtual support systems (e.g., customer support call handling), and Internet messaging systems.

How to find whether how well a Chatbot imitates humans?

Evaluating how well a Chatbot (a computer) imitates humans is a challenging task. To do this, in 1950, Alan Turing, an English Computer Scientist, developed the Turing Test, a method that tests a computer’s ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human (see - https://en.wikipedia.org/wiki/Turing_test). In Turing’s test, he described that the ability of a machine to mimic human behavior can be evaluated by an evaluator who converses with the machine using natural language (text inputs). He designed an experiment that involves a computer (the Chatbot in our case), a human, and an interrogator (evaluator). In his experiment setup, the interrogator cannot see the computer or the human and he continuously post questions to the computer and the human (same questions are posted to both the human and the computer). The interrogator tries to determine the answers generated by the computer after each round of questions. If the interrogator cannot find any differences in the answers provided by the computer and the human, the computer is said to have passed the Turing test. In other words, the machine operates at an intelligence level of a human in answering the questions.

Question 1.

In this exercise, you will interact with several online chatbots. Note that some of them have won the Loebner Prize (i.e., have successfully passed the Turing Test) multiple times. Your task is to converse with the chatbots and evaluate how well they converse with you when asked a standard set of questions (we’ve listed the questions to ask for each of the bots below). Then you will evaluate how each chatbot performed based on the answers you received for the standard questions asked. Record the answers from each chatbot in a table (see the table with questions below). Note that if a chatbot says it does not know the answer to a question you post, try to ask the same question a few times and it may come up with an answer. “I don’t know” is not accepted as a valid answer from any chatbot. If you feel that a chatbot does not know how to answer a particular question after posting the same question a few times, then list the last three responses you received from the chatbot when you post the same question over and over again.

	Bot Name: Mitsuku URL: http://www.square-bear.co.uk/mitsuku/nfchat.htm	Bot Name: Rose URL: http://ec2-54-215-197-164.us-west-1.compute.amazonaws.com/speech.php	Bot Name: Cleverbot URL: http://www.cleverbot.com/
1. How are you?
2. What is your name?
3. What is JavaScript?
4. What is Thanksgiving?

Question 2.

Now, create your own questions for each chatbot and try to see whether the chatbots can still maintain a sensible conversation with you. Ask each chatbot a set of questions (unique for each bot) until the chatbots give you nonsensical answers. Try to get at least one nonsensical answer from each chatbot. Write the questions that you posted for the bot and the nonsensical answer that it came up with, in the table below.

Chatbot	Your Question	Nonsensical Answer
Mitsuku
Rose
Cleverbot

Question 3 Observation:

Write a summary paragraph to record your observations related to how well the bots answered the questions. For example, which type of questions seemed to stump the bots? Which type of questions could they easily answer?

Artificial Intelligence (AI)

Professor John McCarthy, who is named as the father of Artificial Intelligence (AI) first introduced the term Artificial Intelligence at a workshop at Dartmouth College in 1956 (see - https://en.wikipedia.org/wiki/Artificial_intelligence). He defined Artificial Intelligence as “the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable”.

Question 4

In the earlier section, we learned that the Turing Test can be used to test whether a computer program is intelligent or not. Would you consider a computer program is “intelligent” simply because it can pass the Turing Test? Try to argue on this statement by providing examples. You can read further arguments about Artificial Intelligence, Intelligence Machines and the Chinese Room Argument at http://jmc.stanford.edu/artificial-intelligence/what-is-ai/index.html and http://jmc.stanford.edu/articles/chinese.html.

Exercise 3

This part should not be done during the lab. It will require some thought and writing. Complete it after the Wednesday, November 7th lecture, but submit with this lab.

How to Discern Truth

The Internet has been a part of our everyday life. It provides a medium for us to connect with anyone around the world via technologies such as chat, messaging, and email within seconds. We also rely on the Internet to learn new information. However, not every bit of information on the Internet is valid or truthful. In this exercise, we will look at some information sources published on the web and determine ways to identify whether they are valid or not.

Question 1

Watch the video published by Virgin Atlantic about one of their latest aircrafts in the video titled “Is it a bird, is it a plane? Virgin Atlantic harnesses “flapenergy” with new Dreambird 1417”. Here’s the video link - https://www.youtube.com/watch?v=n_ob-5eC8uw

Do you believe that the technology shown in this video is real and the information presented in the video is true? Why or why not?

Use the information from the EvalutingWebSources.pptx to support your reasoning.

Consider the author, publishing body and the date.

Answer the following questions:

Is the information correct or is it propaganda, misinformation or disinformation? Explain.

Document all sources used to support your response.

Question 2

In this exercise, we will experiment with the Google Search engine to determine if it lists correct information in its search results. Before we move to the experiment section, answer the following question.

a. Google search can find any information that we request and it has access to billions of information available on the Internet. Do you think Google search is intelligent as it can find information on the web and can answer questions we ask it? ______________(yes or no and why)

Now go to https://www.google.com and search “is pluto still a planet” (do not use the quotes in your search query). You will see a set of search results presented to you. You may see an article listed in your results titled “Pluto Has Been Officially Reclassified As A Planet! - Futurism”. Read the article titled “Pluto Has Been Officially Reclassified As A Planet! - Futurism” and try to answer the question “is pluto still a planet”. (Link to the Article - https://futurism.com/pluto-reclassified-as-a-major-planet/). Use any additional sources.

b. Do you think that the above article is correct?

Write at least 2 paragraphs explaining your position. Use the information from the EvalutingWebSources.pptx to support your reasoning.

Consider the author, publishing body and the date.

Answer the following questions:

Is the information correct or is it propaganda, misinformation or disinformation? Explain.

Document all sources that used to support your response.

c. If the above article carries false information, do you think Google should not list that article in the search results? Why or why not?

d. Now revisit the statement “Google search is intelligent”. Do you still agree with this statement? Please explain why you would agree or disagree with the above statement.

Rubric:

Exercise		Points
Exercise 1, Part1	Question 1	1
	Question 2	1
	Question 3	1
	Question 4	1
	Question 5	1
	Question 6	1
	Question 7	1
Exercise 1, Part2	Question 1	1
	Question 2	1
	Question 3	1
	Question 4	1
	Question 5	1
	Question 6	1
	Question 7	1
Exercise 2	Question 1	4
	Question 2	3
	Question 3	3
	Question 4	3
Exercise 3	Question 1	3
	Question 2 – a	2
	Question 2 – b	4
	Question 2 – c	2
	Question 2 – d	2
Total		40

9