Artificial Intelligence

Introduction

Dr. Mohsin Ashraf

Assistant Professor
University of Central Punjab, Lahore

Modified slides of CS-188 Berkeley University

Recommended Books
• Artificial Intelligence: A Modern Approach,
Russell & Norvig.
(https://faculty.psau.edu.sa/filedownload/doc-7-pdf-a154ffbcec538a4161a406abf62f5b76-original.pdf )

• “Artificial Intelligence Illuminated”

Ben Coppin,
(http://futuresoft.yolasite.com/resources/Artificial%20Intelligence%20Illuminated.pdf)
Recommended Books
• “Artificial Intelligence – Structures and Strategies for Complex
problem solving”, George F. Luger, Pearson International Edition, Sixth
edition, 2009.
• “Artificial Intelligence: A new synthesis” Nils Nilsson, Morgan
Kaufmann, 1998
• UC Berkeley CS188 Intro to AI (http://ai.berkeley.edu/home.html)
Grading (Tentative)

ITEM WEIGHTAGE
Quizzes 15 %
Assignments 15 %
Class Participation 5%
Mid Term 20 %
Final Term 45 %
Total 100
Logistics
• Quiz may be un-announced.
• Read the book
• Plagiarism – Write your own Solution
• Deadline of the assignments will not be extended
• No Late submissions will be accepted.
Other Announcements
• Email address:
• mohsin.ashraf@ucp.edu.pk
• Office:
• C-Block, Building A, Office 101
• Course Coordinator
• Dr. Mohsin Ashraf

6
Course Topics
• Part I: Solving Problems by Searching
• Fast search / planning
• Constraint satisfaction
• Heuristics
• Adversarial and uncertain search

• Part II: Knowledge, reasoning, and Planning

• Reasoning agents
• Propositional logic
• Predicate logic
• Knowledge-based systems

• Part III: Machine Learning

• Classification
• Clustering
• Neural Networks
Overview

• What is artificial intelligence?

• What can AI do?

• What is this course?

Question
• Is this mouse intelligence?
• It depends on context
• Route planning (Yes)
• Mathematical problem solving (No)

• Is Google Intelligent?
• Task specific intelligence

• Are humans intelligent?

9
Why AI?
• Two main goals of AI:

• To create useful “smart” programs able to do tasks that would normally

require a human expert

• To understand human intelligence better as we test theories of human

intelligence by writing programs which emulate them
 What is Intelligence?
• Learning: The ability of machines to learn from data and improve their
performance over time without being explicitly programmed.
• Reasoning: The ability of machines to use logic and draw conclusions from
available information.
• Perception: The ability of machines to interpret and understand sensory data,
such as images, video, and speech.
• Natural language processing: The ability of machines to understand and process
human language, including written and spoken communication.
• Creativity: The ability of machines to generate new ideas, concepts, and solutions
to problems.
• Decision-making: The ability of machines to make decisions based on available
information and data.
What is Intelligence?
• Ability of problem solving

• Ability to think, plan and schedule

• Efficient memory and information manipulation

• Ability to tackle ambiguous and fuzzy problems

• Ability to understand and perceive

12
Artificial Intelligence
• “Study of computations that make it possible to perceive, reason
and act.”
• Branch of Science which deals with helping machines to find
solutions to complex problems in a more human-like fashion

• In short, putting human intelligence into machines

Artificial Intelligence
Perceive

What is Understand
Intelligence ??
Remember

Think

Reason

Learn from Experience

Adapt to the new

Situation

Problem Solving
Humans Vs Machines
Humans Machines
• Symbolic calculation • Numeric calculation
• Natural language understanding • Machine Language
• Not very precise • Precise
• Knowledge • Data
• Generalize from examples • Cannot generalize
• Deal with noisy inputs • Cannot deal with noise
 What is AI?
The science of making machines that:

Think like people Think rationally

Cognitive modeling Laws of thought
approach approach

Act like people Act rationally

Turing Test Rational agent
approach approach
What is Artificial Intelligence
 Acting Humanly: Turing Test
• Alan Turing (1950) “Computing machinery and intelligence”.

A computer passes the test if a human interrogator,

after posing some written questions, cannot tell
whether the written responses come from a person
or from a computer.
Turing Test Capabilities
 Thinking Humanly:
The cognitive modeling approach
• For that we must determine how humans think.
• Need to get inside the actual workings of human minds.
Three ways to do this:
• Through introspection
• trying to catch our own thoughts as they go by…
• Through psychological experiments
• observing a person in action
• Through brain imaging
• observing the brain in action.
 Thinking Humanly:
The cognitive modeling approach
• Requires scientific theories of internal activities of the brain.
• Cognitive Science: Predicting and testing behavior of human subjects (Hypothesis
based)
• Cognitive Neuroscience: Direct identification from neurological data (Data driven)
• However:
• the available theories do not explain anything resembling human-level general
intelligence
• It may be impossible to identify the detailed structure of human problem solving
using only externally-available data.
21
 Thinking Rationally: Laws of Thought
• Aristotle: First to codify “right thinking”
• Several Greek schools developed various forms of logic:
• Notation and rules of derivation for thoughts
• Propositional logic, Predicate logic, etc.

• Always yield correct conclusions when given correct premises

• E.g. “Socrates is a man; all men are mortal; therefore, Socrates is mortal.”

• These laws of thought were supposed to govern the operation of the mind;
their study initiated the field called logic.
 Thinking Rationally: Laws of Thought
• Problems:
• Not easy to state informal knowledge in logical notation
• Mostly work for toy examples
• Big difference between solving a problem "in principle" and solving it “in
practice”
• Problems with just a few hundred facts can exhaust the computational resources of
any computer
• How to measure if a program is thinking rationally
• There is a difference between thinking and acting
Conclusion:
Right Thinking
Must have rules and logic
Complex and Hard
23
 Acting rationally: Rational agent
• An agent is something that acts.

• All computer programs do something, but

computer agents are expected to do more:
• Operate autonomously,
• Perceive their environment,
• Persist over a prolonged time period
• Adapt to change, and create and pursue goals.
• Make correct inferences

Curiosity (rover)
24
 Acting rationally: Rational agent
• A rational agent is one that acts so as to achieve the best outcome
or, when there is uncertainty, the best expected outcome.
• Rational behavior: doing the right thing
• The right thing: the optimal (best) thing that is expected to maximize
the chances of achieving a set of goals, in a given situation
• Making correct inferences is part of being a rational agent
• There are also ways of acting rationally that cannot be said to involve
inference -- Reflex actions.

25
 Acting rationally: Rational agent
• Advantages over other approaches

• More general than the "laws of thought" approach

• More obedient to scientific development than are approaches based on

human behavior or human thought

• Standard of rationality (Logic) is mathematically well defined and completely

general

• Conclusion:
• Behave in right manner
• Give max. Performance
• Give Optimal Solution 26
 This course is about designing
rational/intelligent agents

Build REPRESENTATIONS that

support MODELS targeted at
THINKING PERCEPTION ACTION
27
Designing Rational Agents
• An agent is an entity that perceives and acts.
• A rational agent selects actions that maximize
its (expected) utility.
• Characteristics of the percepts, environment,
and action space dictate techniques for
selecting rational actions
• This course is about:
• General AI techniques for a variety of

Environment
problem types Sensors
• Learning to recognize when and how a

Agent
Percepts
new problem can be solved with an ?
existing technique
Actuators
Actions
Rational Decisions
• We’ll use the term rational in a very specific, technical way:
• Rational: maximally achieving pre-defined goals
• Rationality only concerns what decisions are made (not the thought
process behind them)
• Goals are expressed in terms of the utility of outcomes
• Being rational means maximizing your expected utility
Approaches to AI
• Strong AI
• Strong AI aims to build machines that can truly reason and solve problems who is self
aware and whose overall intellectual ability is indistinguishable from that of a human
being.
• Human-Like
• Non-human Like

• Weak AI
• Weak artificial intelligence (weak AI), also known as narrow AI, is artificial
intelligence that is focused on one narrow task.
• Deals with the creation of some form of computer-based artificial intelligence that
cannot truly reason and solve problems, but can act as if it were intelligent.
• Siri is a good example of narrow intelligence.
Weak and Strong AI
• Weak AI is AI that can not 'think', i.e. a computer chess playing AI
does not think about its next move, it is based on the programming it
was given, and its moves depend on the moves of the human
opponent.
• Strong AI is the idea/concept that we will one day create AI
that can 'think' i.e. be able to play a chess game that is not based on
the moves of the human opponent or programming, but based on the
AI's own 'thoughts' and feelings and such, which are all supposed to
be exactly like a real humans thoughts and emotions and stuff.
Applications of AI
• Game playing
• Computer vision
• Natural language processing
• Diagnosis systems
• Control
• Optimisation
• Robotics
Natural Language
• Speech technologies (e.g. Siri)
• Automatic speech recognition (ASR)
• Text-to-speech synthesis (TTS)
• Dialog systems

• Language processing technologies

• Question answering
• Machine translation

• Web search
• Text classification, spam filtering, etc…
Vision (Perception)
• Object and face recognition
• Scene segmentation
• Image classification
Robotics
• Robotics

• Technologies
• Vehicles
• Rescue
• Soccer!
• Lots of
automation…
Game Playing
• Classic Moment: May, '97: Deep Blue vs. Kasparov
• First match won against world champion
• “Intelligent creative” play
• 200 million board positions per second
• Humans understood 99.9 of Deep Blue's moves
• Can do about the same now with a PC cluster

• Open question:
• How does human cognition deal with the
search space explosion of chess?
• Or: how can humans compete with computers at all??

• 1996: Kasparov Beats Deep Blue

“I could feel --- I could smell --- a new kind of intelligence across the table.”

• 1997: Deep Blue Beats Kasparov

“Deep Blue hasn't proven anything.”

• Huge game-playing advances recently, e.g. in Go!

Decision Making
• Applied AI involves many kinds of automation
• Scheduling, e.g. airline routing, military
• Route planning, e.g. Google maps
• Medical diagnosis
• Web search engines
• Spam classifiers
• Automated help desks
• Fraud detection
• Product recommendations
• … Lots more!
AI Topics: A Quick Introductory Overview
Classic AI search problems: Map searching (navigation)

39
AI Topics: A Quick Introductory Overview
• Classic AI search problems
• 3*3*3 Rubik’s Cube
Classical AI Search Problems
• 8-Puzzle

1 4 2 1 2 3
6 5 8 4 5 6
7 3 7 8
Classical AI Search Problems
Classical AI Search Problem
• Pac-Man