2
Artificial Intelligence A modern approach (2003)
S. Russell and P. Norvig
=(50%)+(15%)+(35%)
[email protected][email protected]
http:/staff.ustc.edu.cn/~linlixu/ai2013spring/
3
/Introduction and
Agents (chapters 1,2)
/Search (chapters
Review: last chapter
Best-first search
Heuristic functions estimate costs of shortest paths
Good heuristics can dramatically reduce search cost
Greedy best-first search expands lowest h
incomplete and not always optimal
A* search expands lowest g+ h
com
Game playing
Chapter 6
Outline
Games
Perfect play
minimax decisions
- pruning
Resource limits and approximate evaluation
Games of chance
Games of imperfect information
Games vs. search problems
Unpredictable opponent solution is a strategy specifyi
Bayesian networks
Frequentist vs. Bayesian
vs.
Frequentist: probability is the long-run
expected frequency of occurrence. P(A) = n/N, where n is the
number of times event A occurs in N opportunities.
0.10.1
?
Bayesian: degree of belief. It is a measur
Uninformed search
Informed search
1
Review: last week
function TREE-SEARCH( problem, fringe) returns a solution, or failure
fringe INSERT(MAKE-NODE(INTIAL-STATE[problem]), fringe)
loop do
if fringe is empty then return failure
node REMOVE-FRONT (fringe)
First-Order Logic
Chapter 8
Last week
Logical agents apply inference to a knowledge base
to derive new information and make decisions
Basic concepts of logic:
syntax: formal structure of sentences
semantics: truth of sentences wrt models
entailment: ne
Learning from Observations
Chapter 18
Outline
Learning agents
Supervised learning
Decision tree learning
K nearest neighbor learningk
Least squares classification
2
Learning
Learning is essential for unknown environments,
i.e., when designer lacks
SupportVectorMachines
Support
Vector Machines
Supervised learning
Supervisedlearning
Supervisedlearning
AnagentormachineisgivenNsensoryinputsD =cfw_x1,x2 .,xN,aswell
asthedesiredoutputsy
p y1,yy2,.yyN,itsgoalistolearntoproducethe
g
p
correctoutputgivenan
A UNIFIED PERSPECTIVE ON
SUPERVISED, UNSUPERVISED
AND SEMI-SUPERVISED
LEARNING
Problem Types
Problem Types
Supervised learning
Regression
Given
input descriptions x 2 Rn and target values y 2 Rk
y
y
y
x
Infer
y
x
x
x
a function f : X ! Y
Usually
k < n
Lecture 20: Support Vector Machines
Outline
Discriminative learning of classiers.
Learning a decision boundary.
Issue: generalization.
Linear Support Vector Machine (SVM) classier.
Margin and generalization.
Training of linear SVM.
Linear Classicati
Introduction to Support Vector Machines
Colin Campbell,
Bristol University
1
Outline of talk.
Part 1. An Introduction to SVMs
1.1. SVMs for binary classication.
1.2. Soft margins and multi-class classication.
1.3. SVMs for regression.
2
Part 2. General ke
Chapter 5 Classification
Classification: Definition
T id
R e fu n d
M a r ita l
S ta tu s
T a x a b le
In c o m e
C heat
1
S in g l e
125K
No
2
No
M a r r ie d
100K
No
3
No
S in g l e
70K
No
4
Yes
M a r r ie d
120K
No
5
No
D iv o r c e d
95K
Yes
6
No
M a
Association Rule Mining
Motivation
Association Rule Mining
()
Association Rule Mining
Given a set of transactions, find rules that will predict the
occurrence of an item based on the occurrences of other
items in the transaction
Market-Basket transactions
1
Solving Problem By Searching
Chapter 3
http:/staff.ustc.edu.cn/~linlixu/ai2013spring/
Housekeeping
2
Course slides available at
http:/staff.ustc.edu.cn/~linlixu/ai2013spring/
http:/staff.ustc.edu.cn/~linlixu/ai2013spring/
Last chapter
3
Agents interact