Download CSE 214: Data Structures for Information Systems

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Document related concepts

Chinese room wikipedia, lookup

Intelligence explosion wikipedia, lookup

Agent (The Matrix) wikipedia, lookup

Existential risk from artificial general intelligence wikipedia, lookup

AI winter wikipedia, lookup

Visual Turing Test wikipedia, lookup

Wizard of Oz experiment wikipedia, lookup

Alan Turing wikipedia, lookup

Ethics of artificial intelligence wikipedia, lookup

Human–computer interaction wikipedia, lookup

Embodied cognitive science wikipedia, lookup

Knowledge representation and reasoning wikipedia, lookup

Turing test wikipedia, lookup

History of artificial intelligence wikipedia, lookup

Philosophy of artificial intelligence wikipedia, lookup

Transcript
CSE 630: Artificial Intelligence I
Chapter 1: Intro to AI
Jeremy Morris
Spring 2012
1
What would the perfect AI system do?
System that
thinks like
humans
System that
thinks rationally
System that
acts like
humans
System that acts
rationally
“Rational” vs. “Human-like”
“Thinking” vs. “Acting”
2
Acting Humanly: The Turing Test

Test designed by Alan Turing in 1950

Human interrogator interacts with two subjects (A & B)



One human, one machine
“Can you tell if A or B is human?”
If not, then the machine passes
3
Acting Humanly: The Turing Test

What’s needed to pass the Turing Test?






Natural Language Processing
Knowledge Representation
Automated Reasoning
Machine Learning
Natural Language Generation
And, for the “Total Turing Test”




Speech recognition
Speech synthesis
Computer Vision
Robotics
4
Acting Humanly: The Turing Test

What’s wrong with the Turing Test?

Concerned with output, not process


Not reproducible


Encourages deception, hackery
One person’s machine is another’s human being
Not good for mathematical analysis
5
Thinking Humanly: Cognitive Modeling


Understand and reproduce the human thought
process
Related to “cognitive science”



Often uses similar computational models
Difficult to base AI on cognitive modeling
Many successful AI applications without
cognitive modeling
6
Thinking Rationally: “Laws of Thought”

Prescriptive rather than descriptive



What is the “proper” way of reasoning and thinking
(Logic)
Good for building mechanical reasoners
Problems:


Not everything is easy to formulate logically
Large amounts of computation needed to do complete
inference
7
Acting Rationally

Do the right thing!




First hurdle: What is the right thing?
Sometimes the right thing is not the same as what a
human would do
Given: Some (limited) information
Do: Maximize some goal

Traffic light: Maximize traffic flow given sensor inputs
8
Rational Agents

Agent: Entity that perceives and acts



Rational agents perceive and act rationally
Agents try to achieve goals given input perceptions
(percepts)
Functional abstraction: f: Percept*  Action
percept1
percept2
Reasoning
Sensors
action
Effectors
percept3
…
Knowledge
9
Limited Computation

Sometimes we don’t have the resources



Run out of time
Run out of memory
Limited rationality: Do the right thing given the
constraints
10
Foundations of AI

Philosophy (428 BC - )


Mathematics (c. 800 BC -)






Logic, computability, uncertainty
Economics (1776 - )


Logic, Knowledge Representation
Maximizing payoff, adversarial problems
Neuroscience (1861 - )
Psychology (1879 - )
Computer Engineering (1940 - )
Control Theory (1948 - )
Linguistics (1957 - )
11
Current AI Applications

Games




Chatbots



IBM “Deep Blue” beat chess grandmaster Gary
Kasparov
IBM “Watson” playing Jeopardy
But computers remain pathetically bad at “Go”
Cleverbot
Mimicking pet behavior (SONY AIBO)
Advanced Robotics (Honda’s ASIMO)
12
Current AI Applications

Autonomous vehicle routing

DARPA Automatic Vehicle Challenge




Automatic plane collision avoidance
Google Maps and in vehicle mapping devices
Consumer pattern recognition



Google’s autonomous vehicles
Fraud detection
Supermarket tracking
Spam detection
13
Current AI Applications

Telephone-based human-computer interfaces




Airlines
Customer service routing
Roomba “intelligent vacuum”
Machine translation

Google Translate
14
For Next Time


Read Chapters 1 & 2
Do Homework 1


Due at the start of class Thursday!
Start Homework 2

Programming Assignment


Review Course Notes for:



Will take more time than Homework 1
AI Intro – Chapter 1
Agents & Rationality – Chapter 2
BRING QUESTIONS THURSDAY!
15