Transcript Beautiful Thinking - Bradley University

Introduction to Artificial Intelligence
for Bradley University – CS 521
Anthony (Tony) J. Grichnik
Visiting Scientist to Bradley University
Caterpillar Inc.
Outline

Introduction:
The Clans of Artificial Intelligence (AI)

Logical – Be the Expert

Statistical – Would you like to play a game?

Biological – Solutions…naturally

The Future – Hybrids and more
Copyright 2006, Tony Grichnik ~ All Rights Reserved
Anatomy of an Expert System

Framework


Attributes


How would you describe the problem? Is there a metaphor you can
use? (“This is like <<some other thing we understand well>>…”)
What are the key elements of the problem? What do you need to know
to make a good decision?
States / Variable Types


Continuous – Can take on any value, like a decimal number
Ordinal – Has order but is not continuous, like “low, medium, high.”


Categorical – Does not have an implied order, like “apple, orange,
banana”


Often a good way to reduce the complexity of a continuous variable.
A very common mistake is to incorrectly represent a categorical variable as
an ordinal variable!
Boolean – True / false as in most programming languages.
Copyright 2006, Tony Grichnik ~ All Rights Reserved
Anatomy of an Expert System – Continued
Keys to Success
1.
Before doing anything, if you’re going to make an
expert system…
…find good experts!!!!
2.
Begin with the end in mind. Completion and
maintenance depends on managing complexity.
Complexity =
n
a
a = quantization of the attributes
n = number of attributes in the system
Copyright 2006, Tony Grichnik ~ All Rights Reserved
Anatomy of an Expert System – Continued
Pop Quiz!
1.
Which will get you into trouble faster?
1.
2.
Having too many attributes?
Having too many levels per attribute?
Complexity =
n
a
a = quantization of the attributes
n = number of attributes in the system
Copyright 2006, Tony Grichnik ~ All Rights Reserved
Anatomy of an Expert System – Continued
Pop Quiz!
2.
Of the four variable types, which one is to be avoided
whenever possible in an expert system?
1.
2.
3.
4.
Continuous
Ordinal
Categorical
Boolean
Complexity =
n
a
a = quantization of the attributes
n = number of attributes in the system
Copyright 2006, Tony Grichnik ~ All Rights Reserved
Your Assignment

Build an expert system.


Question: What are you an expert at?
Answer: Pizza?
Let’s call it….PizzaExpert!
Copyright 2006, Tony Grichnik ~ All Rights Reserved
PizzaExpert Requirements

Goal: Help your fellow students order the perfect
pizza for their needs from Peoria-area pizzerias.

Measure of success – You’ll be graded on:




Number of pizza places represented
Number of attributes considered
Ease of use
…but most of all…is it really an expert?!
Copyright 2006, Tony Grichnik ~ All Rights Reserved
Getting started on PizzaExpert

First question: how can we gain expertise?

Where to find the pizzerias?


How to know if they’re any good?



Use the Peoria phone directory yellow pages, look under pizza.
Ask someone! (Or order a pizza…8)
Be sure to ask for a menu!
Next, what framework is appropriate? What should the end solution
feel like?

When you go to a restaurant, the first thing you do is look at the ______.


A) Menu!
So our “framework” could be “like an intelligent menu.”

Can you think of a website with an “intelligent” pizza menu?
Copyright 2006, Tony Grichnik ~ All Rights Reserved
PizzaExpert Attributes

Third, what attributes should we consider? Let’s brainstorm:

About the pizza itself





About the pizza experience




Sizes
Type of crust
Type of sauce
Toppings available
Eat in, delivery or both?
Do they deliver to the Bradley campus or area?
What other extras do they have? (Drinks, side dishes, dessert?)
What else can you think of?

I left one out on purpose…
Copyright 2006, Tony Grichnik ~ All Rights Reserved
PizzaExpert – Writing Rules

One more hint…what decision comes first when you
think about pizza?


That should probably be the first rule you apply.
What’s the first question most pizza places ask when you call?
(Other than “can you please hold?” Doesn’t that drive you crazy!)
Smilie’s Pizza
Copyright 2006, Tony Grichnik ~ All Rights Reserved
Introduction to Artificial Intelligence
for Bradley University – CS 521
Anthony (Tony) J. Grichnik
Visiting Scientist to Bradley University
Caterpillar Inc.