Lecture 1: Introduction to Articial Intelligence
Dr. Roman V Belavkin
BIS3226
Contents
1 Problems in AI and Related Areas
1
2 Questions about Intelligence
2
3 Applications of AI
3
References
3
1
Problems in AI and Related Areas
What is AI?
.science and e
Lecture 9: Linear Models
Dr. Roman V Belavkin
BIS3226
Contents
1 Introduction to Modelling
1
2 Linear Functions of One Variable
2
3 Choosing a Model
5
4 Mean-Square Linear Regression
6
1
Introduction to Modelling
What is a Model?
The word model comes fro
Lecture 10: Multilinear Regression
Dr. Roman V Belavkin
BIS3226
Contents
1 Multivariate Data and Models
1
2 Linear Functions of Multiple Variable
3
3 Example: Credit Score Model
5
4 Conclusions
7
1
Multivariate Data and Models
Data-Driven Models
If there
Questions 14:
Case-Based Reasoning
Roman Belavkin
Middlesex University
Question 1
Briey describe the main principles of a case-based reasoning (CBR) expert
system, its operation process and possible dierences in implementation.
Answer:
A CBR system uses
Questions 15:
Genetic Algorithms
Roman Belavkin
Middlesex University
Question 1
Give an example of combinatorial problem. What is the most dicult in
solving these problems?
Answer: One classical example is the Travelling Salesman problem (TSP),
described
Lecture 2s: Elements of Order Theory
Dr. Roman V Belavkin
BIS3226
Binary Relation
Denition 1 (Binary Relation). on set A is any subset R A A.
Notation (a, b) R or aRb means a is related to b.
Example 2. Let A = cfw_, , be three students, and let likes an
Lecture 2: Choice and Optimisation
Dr. Roman V Belavkin
BIS3226
Contents
1 Choice and Preference Relation
1
2 Utility Function
4
3 Multicriteria Decision Making
5
4 Types and Phases of Decisions
7
References
7
1
Choice and Preference Relation
AI and Decis
Lecture 7: Expectation and Correlation
Dr. Roman V Belavkin
BIS3226
Contents
1 Databases and Random Variables
1
2 Measures of Location
2
3 Measures of Dispersion
2
4 Correlation
4
1
Databases and Random Variables
Databases and Random Variables
Case:
Age
G
Lecture 5: Knowledge Representation
Dr. Roman V Belavkin
BIS3226
Contents
1 Introduction to KR
1
2 Types of KR
2
3 Knowledge Synthesis and Engineering
3
1
Introduction to KR
Human Knowledge
Human knowledge is encoded and communicated in a natural languag
Lecture 1s: Elements of Set Theory
Dr. Roman V Belavkin
BIS3226
Contents
1 Sets and Operations on Sets
1
2 Correspondences and Mappings between Sets
3
The bishop gave monkey the banana.
only
Only the bishop gave the monkey the banana.
The only bishop gave
Lecture 6: Uncertainty and Information
Dr. Roman V Belavkin
BIS3226
Contents
1 Introduction
1
2 What is probability?
2
3 Conditional Probability and Independence
4
4 Uncertainty and Information
6
1
Introduction
Sources of Uncertainty
Complexity : the numb
Lecture 4: Search
Dr. Roman V Belavkin
BIS3226
Contents
1 Introduction to Section Problems
1
2 Types and Examples of Search Strategies
3
3 Search in Rule-Based Systems
5
1
Introduction to Section Problems
A Simple Search Problem
Problem 1 (Choice). Choose
Lecture 3: Logic and Rule-Based Reasoning
Dr. Roman V Belavkin
BIS3226
Contents
1 Introduction to Production Systems
1
2 Elements of Boolean Logic
2
3 Problem Solving
5
4 Development and Operation of ES
6
5 Discussion
9
1
Introduction to Production System
Lecture 14: Case-Based Reasoning
Dr. Roman V Belavkin
BIS3226
Contents
1 Case-Based Reasoning
1
2 Operation of CBR Systems
3
3 Discussion, Examples and Applications
4
1
Case-Based Reasoning
Case-Based Reasoning Systems
Instead of facts and rules, CBR sys
Lecture 11: Feed-Forward Neural Networks
Dr. Roman V Belavkin
BIS3226
Contents
1 Biological neurons and the brain
1
2 A Model of A Single Neuron
3
3 Neurons as data-driven models
5
4 Neural Networks
6
5 Training algorithms
8
6 Applications
10
7 Advantages
Questions 2:
Choice and Optimisation
Roman Belavkin
Middlesex University
Question 1
Consider a set of all integers z such that z 2 < 10. Is it an ordered set? Does
this set have the top (maximum) or the bottom (minimum) elements? What
are their values if
Questions 3:
Logic and Rule-Based Reasoning
Roman Belavkin
Middlesex University
Question 1
Use the equivalence of Boolean operations and set-theoretic operations to
prove the duality (De Morgans) laws:
(a b) = a b ,
(a b) = a b
Hint: you can use Venn diag
Questions 6:
Uncertainty and Information
Roman Belavkin
Middlesex University
Question 1
Name and briey describe three main sources of uncertainty.
Answer: Here, we emphasise the following sources of uncertainty: Complexity, ignorance and randomness.
Compl
Questions 5:
Knowledge Representation
Roman Belavkin
Middlesex University
Question 1
Describe four approaches to knowledge representation according to Mylopoulos & Levesque (1984).
Answer:
Logical calculus is used, such as rst-order predicate calculus, mo
Questions 4:
Search
Roman Belavkin
Middlesex University
Question 1
Briey describe what is uninformed and informed search? Name examples
of uninformed and informed search strategies in lists and trees.
Answer:
Uninformed search uses some xed strategy that
Questions 8:
Game Theory
Roman Belavkin
Middlesex University
Question 1
Suppose of you have a choice between two lotteries A and B:
Lottery A: The utility can have values -1, 0 or 1.
Lottery B: The utility can have values -2, 0 or 2.
Suppose that all va
Questions 7:
Expectation and Correlation
Roman Belavkin
Middlesex University
Question 1
Let variable x can have values 1, 2 and 3 with probabilities P (1) = 1/5,
P (2) = 3/5 and P (3) = 1/5. What is the expected value of x? Compare it
with mean value of (
Questions 10:
Multilinear Regression
Roman Belavkin
Middlesex University
Question 1
Suppose that a multiple regression model using m = 6 input variables should
be created based on some data. What is the minimum number of cases the
dataset should have?
Ans
Questions 9:
Linear Models
Roman Belavkin
Middlesex University
Question 1
What is a model and its error, and how can they be dened in terms of
information?
Answer: A model is a simplied representation of another object. Because the model contains fewer de
Questions 13:
Self-Organising maps
Roman Belavkin
Middlesex University
Question 1
Below is a diagram of a selforganising map:
A
B
C
D
E
r
B
b
U o &
&
r Tt
0
T b 0
t rrt T& &
t
t &
&
rr
t
r &
t &
r r
&
&
t
t
t &t r&
&
& rr
t&
&
t
r
Questions 11:
Feed-Forward Neural Networks
Roman Belavkin
Middlesex University
Question 1
Below is a diagram if a single articial neuron (unit):
x1
w1
~
w2 E
E
v
w&
3&
b
&
&
x2
y = (v)
x3 &
Figure 1: Single unit with three inputs.
The node has three i
Lecture 12: Clustering
Dr. Roman V Belavkin
BIS3226
Contents
1 Metric Spaces
1
2 Data as Vectors in Metric Spaces
3
3 The Clustering Problem
4
1
Metric Spaces
Metric Spaces
Let X be a set. How can we compare the elements of X?
Denition 1 (Metric). is a fu
Lecture 8: Game theory
Dr. Roman V Belavkin
BIS3226
Contents
1 Expected Utility and Decision-Making under Uncertainty
1
2 Games
3
3 Finite Zero-Sum 2-Person Games
3
4 Mixed Strategies
5
Historical Background
1654 Blaise Pascal and Pierre Fermat
1657 Chris