CMSC 100 Artificial Intelligence: Human vs. Machine Professor Marie desJardins

advertisement
CMSC 100
Artificial Intelligence: Human vs. Machine
Professor Marie desJardins
Thursday, November 8, 2012
Artificial Intelligence
1
Thu 11/1/12
Memory is at the Core
(Literally)


Remember Hal?

“Open the pod bay door, Hal.”

“My mind is going...”
Memory is at the core of our being (and a computer’s)
The first magnetic core memory [www.columbia.edu/acis/history]
Artificial Intelligence
2
thebrain.mcgill.ca
Thu 11/1/12
Overview

What is AI? (and why is it so cool?)

AI: Past and Present


History of AI
AI Today

Computational vs. Biological Memory

The Skeptics Speak
Artificial Intelligence
3
Thu 11/1/12
What is AI?
Artificial Intelligence
4
Thu 11/1/12
AI: A Vision

Could an intelligent agent living on your home computer manage your
email, coordinate your work and social activities, help plan your
vacations…… even watch your house while you take those well planned
vacations?
5
Main Goals of AI

Represent and store knowledge

Retrieve and reason about knowledge

Behave intelligently in complex environments

Develop interesting and useful applications

Interact with people, agents, and the environment
Artificial Intelligence
6
Thu 11/1/12
Foundations of AI
Mathematics
Economics
Psychology
Computer
Science &
Engineering
AI
Cognitive
Science
7
Philosophy
Biology
Linguistics
Big Questions






Artificial Intelligence
Can machines think?
If so, how?
If not, why not?
What does this say about human beings?
What does this say about the mind?
And if we can make machines think,
should we?
8
Thu 11/1/12
History of AI
Artificial Intelligence
9
Thu 11/1/12
Timeline

1950: Alan Turing proposes the Turing test

1956: Dartmouth conference, “artificial intelligence” coined

1958: Lisp programming language for symbolic reasoning created

1965: First expert system (Dendral – chemistry modeling)

1969: SRI’s Shakey the robot: perception and problem solving

1971: BKG backgammon program defeats the world champion

1997: Deep Blue defeats Garry Kasparov, world chess champion

2007: Checkers is solved (unbeatable program created)

2012: Watson beats Jeopardy! champions Ken Jennings and Brad Rutter
Artificial Intelligence
10
Thu 11/1/12
Early AI: Eliza

ELIZA: A program that simulated a psychotherapist interacting
with a patient and successfully passed the Turing Test

Implemented at MIT during 1964-1966 by Joel Weizenbaum

First script was DOCTOR



Simple pattern-matching
Each pattern had an associated reply which might include bits of the
input (after simple transformations (my  your)
Weizenbaum was shocked at reactions



Artificial Intelligence
Psychiatrists thought it had potential
People unequivocally anthropomorphized
 Let me tell you my problems...
Many thought it solved the natural language problem!!
11
Thu 11/1/12
Eliza Speaks
I am the psychotherapist. Please, describe your problems.
i'm having a hard time with my students
Why do you say that?
well, they keep asking me for money
Maybe your plans have something to do with this.
what plans?
Is it because of your life that you say what plans?
No, it's because I didn't know what you were talking about
Is it because you didn't know what I was talking about that you came to me?
no, it's because i wanted to see if you could pass the turing test
Does it bother you that you wanted to see if I could really pass the turing
test?
…..
How do you reconcile problems at school?
T his is a waste of time. Good bye
My secretary will send you a bill.
Artificial Intelligence
12
Thu 11/1/12
Early AI: Parry

Kenneth Colby modeled a paranoid using the same techniques as
Eliza, around 1968

PARRY has basic emotions. If it gets angry, its replies become more
hostile

In a Turing test, expert psychiatrists were unable to distinguish
PARRY’s ramblings from those of real paranoids
Artificial Intelligence
13
Thu 11/1/12
PARRY Speaks
Good afternoon. My name is Doctor Jones.
Pleased to meet you.
How are you?
All right I guess.
Could you tell me your name?
Pat Smith.
I though you might have some problems.
I've been upset lately.
By what?
People get on my nerves sometimes.
Tell me more.
Do you know anything about bookies?
...
Artificial Intelligence
14
Thu 11/1/12
P: I don't understand your motives.
E: Why don't you understand my motives?
P: Shouldn't I be?
E: Why do you say that just now?
P: That's just the way it is.
E: Suppose that's the way it were not.
P: So?
E: I am not sure I understand you fully.
P: You harp on the same questions too much.
E: What does that similarity suggest to you?
P: I don't get you.
E: Do you wish that you did get I?
Artificial Intelligence
15
Thu 11/1/12
Turing Test

Three rooms contain a person, a computer, and an interrogator

The interrogator can communicate with the other two by
“teleprinter” (or, say, AIM)

The interrogator tries to determine which is the person and which is
the machine

The machine tries to fool the interrogator into believing that it is the
person

If the machine succeeds, then we conclude that the machine can
think
Artificial Intelligence
16
Thu 11/1/12
The Loebner Contest

A modern version of the Turing Test, held annually, with
a $100,000 cash prize




Participants include humans, computers (chatbots), and
judges
Scoring




Hugh Loebner was once director of UMBC’s Academic Computing
Services (née UCS, lately OIT)
http://www.loebner.net/Prizef/loebner-prize.html
Rank from least human to most human
Highest median rank wins $2000
If better than a human, win $100,000 (Nobody yet…)
2012 winner: Chip Vivant, http://www.chipvivant.com/
Artificial Intelligence
17
Thu 11/1/12
What’s Easy and What’s
Hard?

It’s been easier to mechanize many of the high-level tasks we usually
associate with “intelligence” in people
 e.g., symbolic integration, proving theorems, playing chess,
medical diagnosis

It’s been very hard to mechanize tasks that lots of animals can do


walking around without running into things

catching prey and avoiding predators

interpreting complex sensory information (e.g., visual, aural, …)

modeling the internal states of other animals from their behavior

working as a team (e.g., with pack animals)
Is there a fundamental difference between the two categories?
Artificial Intelligence
18
Thu 11/1/12
AI Today
Artificial Intelligence
19
Thu 11/1/12
Who Does AI?

Academic researchers (perhaps the most Ph.D.-generating area of
computer science in recent years)


Government and private research labs


Some of the top AI schools: CMU, Stanford, Berkeley, MIT, UIUC,
UMd, U Alberta, UT Austin, ... (and, of course, UMBC!)
NASA, NRL, NIST, IBM, AT&T, SRI, ISI, MERL, ...
Lots of companies!
Artificial Intelligence
20
Thu 11/1/12
Computational vs. Biological
Memory
Artificial Intelligence
21
Thu 11/1/12
How Does It Work?
(Humans)

Basic idea:



Chemical traces in the neurons of the brain
Types of memory:

Primary (short-term)

Secondary (long-term)
Factors in memory quality:

Distractions

Emotional cues

Repetition
Artificial Intelligence
22
Thu 11/1/12
How Does It Work?
(Computers)



Basic idea:

Store information as “bits” using physical processes (stable electronic
states, capacitors, magnetic polarity, ...)

One bit = “yes or no”
Size
Types of computer storage:

Primary storage (RAM or just “memory”)

Secondary storage (hard disks)

Tertiary storage (optical jukeboxes)

Off-line storage (flash drives)
Speed
Factors in memory quality:

Power source (for RAM)

Avoiding extreme temperatures
Artificial Intelligence
23
Thu 11/1/12
Measuring Memory

Remember that one yes/no “bit” is the basic unit

Eight (23) bits = one byte

1,024 (210) bytes = one kilobyte (1K)*

1,024K (220 bytes) = one megabyte (1M)

1,024K (230 bytes) = one gigabyte (1G)

1,024 (240 bytes) = one terabyte (1T)

1,024 (250 bytes) = one petabyte (1P)

... 280 bytes = one yottabyte (1Y?)
* Note
Artificial Intelligence
that external storage is usually measured in decimal rather than binary (1000 bytes = 1K, and so on)
24
Thu 11/1/12
Showdown

Computer capacity:








Primary: 10-7 sec.
Secondary: 10-5 sec.

Computing capacity: 1 petaflop
(1015 floating-point instructions per
second), very special purpose
Digital
Extremely reliable
Not (usually) parallel
Artificial Intelligence
Human capacity:

Computer retrieval speed:




Primary storage: 64GB
Secondary storage: 750GB (~1012)
Tertiary storage: 1PB? (1015)
25
Primary storage: 7 ± 2
“chunks”
Secondary storage: 108432 bits??
(or maybe 109 bits?)
Human retrieval speed:


Primary: 10-2 sec
Secondary: 10-2 sec

Computing capacity: possibly
100 petaflops, very general
purpose



Analog
Moderately reliable
Highly parallel
Thu 11/1/12
It’s Not Just What You
“Know”






Storage
Indexing
Retrieval
Inference
Semantics
Synthesis

...So far, computers are good at storage, OK at indexing and retrieval,
and humans win on pretty much all of the other dimensions

...but we’re just getting started

Electronic computers were only invented 60 years ago!

Homo sapiens has had a few hundred thousand years to evolve...
Artificial Intelligence
26
Thu 11/1/12
The Skeptics
Speak
Artificial Intelligence
27
Thu 11/1/12
Mind and Consciousness

Many philosophers have wrestled with the question:


Is Artificial Intelligence possible?
John Searle: most famous AI skeptic

Chinese Room argument
?
!
Artificial Intelligence
28
Thu 11/1/12
What Searle Argues

People have beliefs; computers and machines don’t.

People have “intentionality”; computers and machines don’t.

Brains have “causal properties”; computers and machines don’t.

Brains have a particular biological and chemical structure; computers
and machines don’t.

(Philosophers sometimes make claims like “People have intentionality”
without actually saying what “intentionality” is, except (in effect) “the
stuff that people have and computers don’t.”)
Artificial Intelligence
29
Thu 11/1/12
Let’s Introspect For a
Moment...



Have you ever learned something by rote that you didn’t really
understand?
Were you able to get a good grade on an essay where you didn’t really
know what you were talking about?
Have you ever convinced somebody you know a lot about something
you really don’t?




Are you a Chinese room??
What does “understanding” really mean?
What is intentionality? Are human beings the only entities that can
ever have it?
What is consciousness? Why do we have it and other animals and
inanimate objects don’t? (Or do they?)
Artificial Intelligence
30
Thu 11/1/12
Just You Wait...
Artificial Intelligence
31
Thu 11/1/12
Download