Download Report

Intorduction to
Artificial Intelligence
Prof. Dechter
ICS 270A
Winter 2003
Course Outline





Classoom: ICS2-144
Days: Tuesday & Thursday
Time: 09:30 a.m. - 10:50 a.m.
Instructor: Rina Dechter
Textbooks



Nils Nilsson, "Artificial Intelligence: A New
Synthesis", Morgan Kauffmann, 1998
S. Russell and P. Norvig, "Artificial
Intelligence: A Modern Approach" (Second
Edition), Prentice Hall, 1995
J. Pearl, "Heuristics: Intelligent Search
Stratagies", Addison-Wesley, 1984.
270a- winter 2003
Course Outline
Assignments:
 There will be weekly homework-assignments, a project, a
midterm and/or a final.
Course-Grade:
 Homeworks plus project will account for 50% of the grade,
midterm and/or final 50% of the grade.
Course Overview
 Topics covered Include: Heuristic search, Adverserial
search, Constraint Satisfaction Problems, knowledge
representation, propositional and first order logic, inference
with logic, Planning, learning and probabilistic reasoning.
270a- winter 2003
Course Outline
Week
Week
1
Topic
Date
Introduction and overview: What is AI? History
7-Jan
Nillson Ch.1 (1.1-1.5), RN: chapters 1,2.
Problem solving: Statement of Search problems: state space graph, problem
types, examples (puzzle problem, n-queen, the road map, travelling salesman.)
Nillson Ch 7. RN: chapter 3, Pearl: ch.1
Week
2
Uninformed search: Greedy search, breadth-first, depth-first, iterative
deepening, bidirectional search.
14-Jan
Nillson Ch. 8, RN: Ch. 3, Pearl: 2.1, 2.2
Informed heuristic search: Best-First, Uniform cost, A*, Branch and bound.
Nillson Ch. 9, RN: Ch. 4 , Pearl, 2.3.1
Week
3
Properties of A*, iterative deepening A*, generating heuristics automatically.
Learning heuristic functions.
28-Jan
Nillson Ch. 9, 10.3, RN: chapter 4, Pearl: 3.1, 3.2.1, 4.1, 4.2
Game playing: minimax search, alpha-Beta pruning.
Nillson Ch. 12, RN: Ch. 6.
270a- winter 2003
Course Outline
Week 4
Constraint satisfaction problems
21-Jan
Definitions, examples, constraint-graph, constraint propagation
(arc-consistency, path-consistency), the minimal network.
Reading: RN: Ch. 5, class notes.
Backtracking and variable-elimination
advanced search: forward-checking, Dynamic variable orderings,
backjumping, solving trees, adaptive-consistency.
Reading: RN: Ch. 5, class notes.
Week 5
Knowledge and Reasoning: Propositional logic, syntax,
semantics, inference rules.
4-Feb
Nillson Ch. 13, RN: Ch 7.
Propositional logic. Inference, First order logic
Nillson Ch. 14, RN: Ch. 7
Week 6
Knowledge representation:
11-Feb
First-order (predicate) Logic.
Nillson Ch. 15, RN: Ch. 9.
270a- winter 2003
Course Outline
Week 7
Inference in First Order logic
18-Feb
Nillson Ch. 16, RN: Ch. 9
Planning:
Logic-based planning, the situation calculus, the frame
problem.
Nillson Ch. 21, RN: Ch. 11.
Week 8
Planning: Planning systems, STRIP, regression planning,
current trends in planning: search-based, and
propositional-based.
25-Feb
Nillson Ch. 22, RN: Ch. 11.
Week 9
Reasoning and planning under uncertainty
4-Mar
Nillson Ch. 19, RN: chapter 14.
Week 10
Assorted topics
11-Mar
270a- winter 2003
Course Outline
Resources on the Internet
 AI on the Web: A very comprehensive list of Web
resources about AI from the Russell and Norvig
textbook.
Essays and Papers
 What is AI, John McCarthy
 Rethinking Artificial Intelligence, Patrick H. Winston
 International Summer School on AI Planning
 An overview of recent algorithms for AI planning, Jussi
Rintanen
270a- winter 2003
Today’s class





What is Artificial Intelligence?
Engineering versus cognitive approaches
Intelligent agents
History of AI
Real-World Applications of AI

many products, systems, have AI
components
270a- winter 2003
What is Artificial
Intelligence?




Thought processes vs behavior
Human-like vs rational-like
RN figure:
“How to simulate humans intellect and
behavior on by a machine.




Mathematical problems (puzzles, games,
theorems)
Common-sense reasoning
Expert knowledge: lawyers, medicine,
diagnosis
Social behavior
270a- winter 2003
What is Artificial Intelligence

Thought processes


“The exciting new effort to make computers
think .. Machines with minds, in the full and
literal sense” (Haugeland, 1985)
Behavior

“The study of how to make computers do
things at which, at the moment, people are
better.” (Rich, and Knight, 1991)
270a- winter 2003
270a- winter 2003
The Turing Test

Requires
Natural language
 Knowledge representation
 Automated reasoning
 Machine learning (vision, robotics)

270a- winter 2003
Acting humanly


Turing test (1950)
Requires:






Thinking humanly:



Introspection, the general problem solver (Newell and
Simon 1961)
Cognitive sciences
Thinking rationally:



Natural language
Knowledge representation
automated reasoning
machine learning
(vision, robotics.) for full test
Logic
Problems: how to represent and reason in a domain
Acting rationally:

Agents: Perceive and act
270a- winter 2003
AI examples
Common sense reasoning


Tweety
Yale Shooting problem
Update vs revise knowledge
The OR gate example: A or B - C
 Observe C=0, vs Do C=0
Chaining theories of actions

Looks-like(P)  is(P)
Make-looks-like(P)  Looks-like(P)
---------------------------------------Makes-looks-like(P) ---is(P) ???
Garage-door example: garage door not included.
 Planning benchmarks
 8-puzzle, 8-queen, block world, grid-space world
 (Nillson Fig 1.2)
270a- winter 2003
History of AI

McCulloch and Pitts (1943)


Minsky (1951)


Built a neural net computer
Darmouth conference (1956):




Neural networks that learn
McCarthy, Minsky, Newell, Simon met,
Logic theorist (LT)- proves a theorem in Principia
Mathematica-Russel.
The name “Artficial Intelligence” was coined.
1952-1969






GPS- Newell and Simon
Geometry theorem prover - Gelernter (1959)
Samuel Checkers that learns (1952)
McCarthy - Lisp (1958), Advice Taker, Robinson’s
resolution
Microworlds: Integration, block-worlds.
1962- the perceptron convergence (Rosenblatt)
270a- winter 2003
History, continued

1966-1974 a dose of reality


Problems with computation
1969-1979 Knowledge-based systems


Weak vs. strong methods
Expert systems:
• Dendral:Inferring molecular structures
• Mycin: diagnosing blood infections
• Prospector: recomending exploratory drilling (Duda).


1980-1988: AI becomes am industry




Roger Shank: no syntax only semantics
R1: Mcdermott, 1982, order configurations of computer
systems
1981: Fifth generation
1986-present: return to neural networks
Recent event:

Hidden markov models, planning, belief network
270a- winter 2003
What’s involved in Intelligence?
Intelligent agents

Ability to interact with the real world





Knowledge Representation, Reasoning and
Planning




to perceive, understand, and act
e.g., speech recognition and understanding and
synthesis
e.g., image understanding
e.g., ability to take actions, have an effect
modeling the external world, given input
solving new problems, planning and making decisions
ability to deal with unexpected problems, uncertainties
Learning and Adaptation


we are continuously learning and adapting
our internal models are always being “updated”270a- winter 2003
• e.g. a baby learning to categorize and recogniz animals
Implementing an Agent
270a- winter 2003
Implementing agents


Table look-ups
Autonomy




All actions are completely specified
no need in sensing, no autonomy
example: Monkey and the banana
Structure of an agent


agent = architecture + program
Agent types
•
•
•
•
•
•
medical diagnosis
Satellite image analysis system
part-picking robot
Interactive English tutor
cooking agent
taxi driver
270a- winter 2003
Agent types


Example: Taxi driver
Simple reflex


Agents that keep track of the world



If car-in-front-is-breaking and on fwy then initiatebreaking
needs internal state
goal-based



If car-in-front-is-breaking then initiate-breaking
If car-in-front-is-breaking and needs to get to hospital
then go to adjacent lane and plan
search and planning
utility-based


If car-in-front-is-breaking and on fwy and needs to
get to hospital alive then search of a way to get to the
hospital that will make your passengers happy.
270a- winter 2003
Needs utility function that map a state to a real
function (am I happy?)
AI Application: Reasoning

Scheduling:




Puzzle solving






Nasa Space telescope
factory scheduling
class scheduling
Chess
Checkers
Backgamon
Speech recognition
Vision
Diagnosis




Medical
Circuit diagnosis
Health care consulting
Decision support systems
270a- winter 2003
Summary of State of AI Systems , Practice

Speech synthesis, recognition and understanding



Computer vision



adaptive systems are used in many applications: have their
limits
Planning and Reasoning



works for constrained problems (hand-written zip-codes)
understanding real-world, natural scenes is still too hard
Learning


very useful for limited vocabulary applications
unconstrained speech understanding is still too hard
only works for constrained problems: e.g., chess
real-world is too complex for general systems
Overall:


many components of intelligent systems are “doable”
there are many interesting research problems remaining
270a- winter 2003
Summary

What is Artificial Intelligence?



History of AI
Intelligent agents


modeling humans thinking, acting, should think,
should act.
We want to build agents that act rationally
Real-World Applications of AI

AI is alive and well in various “every day” applications
• many products, systems, have AI components

Assigned Reading


Chapter 1, Nillson
Chapters 1 and 2 in the text R&N
270a- winter 2003