Course : Artificial Intelligence

Effective Period : July 2020

Introduction to Artificial
Intelligence

Session 01

1
 Learning Outcomes
At the end of this session, students will be able to:
 LO 2: Describe what is AI and identify concept of intelligent
agent

2
 Outline
1. Artificial Intelligence (AI)
2. Foundations of AI
3. Intelligent Agent
4. Concept of Rationality
5. Properties of Environment
6. Structure of Intelligent Agents

3
 Artificial Intelligence

Artificial intelligence (AI) attempts not just to understand but

also to build intelligent entities

Stuart J Russel, Professor in Univ. of California Berkeley

Peter Norvig, Director of Research

4
 Artificial Intelligence

Just as the Industrial Revolution freed up a lot of humanity from

physical drudgery, I think AI has the potential to free up humanity

from a lot of the mental drudgery [source]

Andrew Ng, Adjunct Professor in Stanford

5
 Artificial Intelligence

AI is going to bring major shifts in society through developments

in self-driving cars, medical image analysis, better medical

diagnosis, and personalized medicine [source]

Yann LeCun, Director of AI Research, Facebook

6
 What is AI?
Thinking Humanly Thinking Rationally
“The exciting new effort to make “The study of mental faculties
computers thinks … machine through the use of
with minds, in the full and literal computational models”
sense” (Charniak and McDermott, 1985)
(Haugeland,
Acting 1985)
Humanly Acting Rationally
“The study of how to make “Computational Intelligence is
computers do things at which, at the study of the design of
the moment, people are better” intelligent agents”
(Rich and Knight, 1991) (Poole et al., 1998)

7
 Acting humanly
• The Turing Test (Alan Turing, 1950)

• Instead of duplicating an exemplar, it is more useful and

important to study the principles of the intelligence

8
 Thinking humanly
• The cognitive modelling
– Part of cognitive science
• Once we have a sufficiently precise theory of the mind, it
becomes possible to express the theory as a computer
program

9
 Thinking rationally
• The “laws of thought” or logic

• Aristotle was one of the first to introduce logic or “right

thinking”
• Syllogisms provided patterns for argument structure
– Correct premises = correct solutions

– Example:
• He is a boy; All boys are handsome;

• Therefore, he is handsome

10
 Acting rationally
• Rational behavior: doing the right thing
• The right thing: that which is expected to maximize goal
achievement given the available information
– Example:
• Given money Rp. 100.000, you have to buy a birthday
gift
• Which one will you choose? Which one is more
rational?
– Bracelet (Rp. 100.000) or Ring (Rp. 150.000)
11
 Rational Agent
• An agent
– Just something that acts
• A rational agent
– One that acts as to achieve the best (expected)
outcome
• The rational-agent approach has two advantages:
– It is more general than the “laws of thought” approach
– it is more amenable to scientific development

12
 AI Applications
• Speech Recognition
– Virtual Assistants
• Siri (Apple)
• Google Now
• Cortana (Microsoft)
– They helps us to arrange
[source]
meetings, check weather,
do a phone call, send a
message, etc.
13
 AI Applications
• Machine Translation
– Google Translate
– No more bringing dictionary when travelling

[source]
14
 AI Applications
• Robotics
– Several robots were sent to Fukushima nuclear tragedy
to perform various tasks (Left)
– Hubo: A KAIST Robot who wins the DARPA Robot
Challenge (Right)

[source] [source] 15
 AI Applications
• Recommendation Systems
– AI helps to provide items / photos / various things based
on our social activities
– Instagram Explore / Search Feed
• If we follow many badminton accounts, they show:

16
 AI Applications
• Search Engines
– Google’s search engines algorithm is designed to show
internet pages of our interests in a blink of eyes

17
 AI Applications
• Email
Sender:
– Spam/Junk email detection • [email protected]
• [email protected]
• [email protected]
Title:
• Hi, I need your help!
• 1 million dollar for you now!
• [blank]
Content:
• Congrats, you won XXX
Awards. Please tell me your
name, address, birth date, and
telephone number to
[email protected]
18
 AI Applications
• Face Detection

[source]
19
 AI Applications
• Face Recognition
– China’s facial recognition technology to identify the
citizen

[source] 20
 AI Applications
• Games
– Chess (1997): Kasparov vs. IBM Deep Blue

• Powerful search algorithms

– Jeopardy! (2011): Humans vs. IBM Watson

• Natural language processing and information extraction

– Go (2016): Lee Sedol vs. Google AlphaGo

• Deep learning + reinforcement learning + search

algorithms!
– Future: 5 vs 5 AI for Dota
21
 AI Applications
• Games

– March 2016, AlphaGo beat Lee Sedol (4 vs 1)

– December 2016, AlphaGo beat Ke Jie (3 vs 0)

[source] [source]
22
 AI Applications
• Autonomous Driving
– NuTonomy: A robo-taxi service in Singapore

[source]
23
 AI Applications
• Autonomous Driving
– Waymo: An autonomous car company under Alphabet,
Inc. (Google’s parent company)

[source]
24
 Foundation of AI

Philosophy

Linguistics Mathematics

Control
Theory and
Cybernetics AI Economics

Computer
Neuroscience
Engineering

Psychology

25
 Foundation of AI
• Philosophy

– Logic, methods of reasoning

– Foundations of learning, language, rationality

• Mathematics
– Logic: Formal representation and proof

– Algorithms, computation, (un)decidability, (in)tractability

– Probability

26
 Foundation of AI
• Economics
– Formal theory of rational decisions
• Neuroscience
– Plastic physical substrate for mental activity
• Psychology
– Adaptation
– Phenomena of perception and motor control
– Experimental techniques (psychophysics, etc.)

27
 Foundation of AI
• Computer Engineering
– How can we build an efficient computer to build AI
program?
• Control Theory and Cybernetics
– Simple optimal agent designs
• Linguistics
– Knowledge representation
– Grammar

28
Brief History of AI

29
 Agent
• An agent is anything that can be viewed as perceiving its
environment through sensors and acting upon that
environment through actuators.

30
 Agent

A vacuum-cleaner world with just two locations

Percept: location and contents, i.e. [A, Dirty]
Actions: Left, Right, Suck, NoOp

31
 Agent
Partial tabulation of the simple agent function

32
 Concept of Rationality
Definition of a rational agent:
• For each possible percept sequence, a rational agent
should select an action that is expected to maximize its
performance measure, given the evidence provided by
the percept sequence and whatever built-in knowledge
the agent has.

33
 Task Environments
To build a rational agent, we need to first define the PEAS:
• P (Performance)
• E (Environment)

• A (Actuators)
• S (Sensors)

34
 What is PEAS for?
• Autonomous car

[source]

– Performance: safety, destination, legality, comfort

– Environment: streets, pedestrian, highway, weather

– Actuators: steering, accelerator, brake, horn

– Sensors: video, GPS, accelerometer, keyboard

35
 Structure of Intelligent Agents
• Agent = Architecture + Program
– Architecture is the hardware
• Sensors + Actuators
– Program is the software

Progra
m

36
 Agent Types

Four basic type of agents:

• Simple reflex agents

• Modal-based reflex agents

• Goal-based agents

• Utility-based agents

All of which can be generalized into learning agents that can

improve their performance and generate better actions.
37
 Simple Reflex Agents
An action is done based on the current state only. Ignore the sensors
history.

38
 Modal-based Reflex Agents
The sensors and actions history is used to model the world /
environment. An action is done based on the world model.

39
 Goals-based Agents
An action is done based on the combined information from the
world model and goal information.

40
 Utility-based Agents
An action is done based on the agent happiness (utility).
It is the agent’s performance measure.

41
 Learning Agents
Programming agents by hand can be very tedious. Some more
expeditious method seem desirable" Alan Turing, 1950.

42
 References
• Stuart Russell, Peter Norvig. 2010. Artificial Intelligence :
A Modern Approach. Pearson Education. New Jersey.
ISBN:9780132071482

43