ARTIFICIAL
INTELLIGENCE
Chapter 1: Introduction
Instructor: Dr Ghulam Mustafa
Outline
• What is Artificial Intelligence (AI)?
• The foundations of AI
• The history of AI
• The state of the art
Artificial Intelligence
Artificial Intelligence
• Homo sapiens – man the wise
• How we think i.e.
– Perceive, understand, predict & manipulate
• AI is not just to understand but build
intelligent entities -- Machines that act
effectively and safely in novel situations.
• AI one of the most interesting and fastest-
growing fields
What is AI?
"The study of agents that receive percepts from the
environment and perform actions to maximize their
success."
They categorize AI into four approaches:
1. Thinking humanly – Modeling human cognition.
2. Thinking rationally – Using logic and reasoning.
3. Acting humanly – Mimicking human behavior.
4. Acting rationally – Making optimal decisions based on
goals.
The "Acting Rationally" approach, where AI is seen as an
intelligent agent that perceives and acts to achieve the
best outcome, is the most widely accepted definition.
Acting humanly: Turing Test
• The Turing Test, proposed by Alan Turing (1950)
• "Can machines think?" → "Can machines behave intelligently?"
• Operational test for intelligent behavior: the Imitation Game
• Predicted that by 2000, a machine might have a 30% chance of
fooling a lay person for 5 minutes
• Suggested major components of AI: knowledge, reasoning,
language understanding, learning
• Total Turing test with video signal: computer vision, robotics
Thinking humanly: cognitive
modeling
• Requires scientific theories of internal activities
of the brain
• How to validate?
– introspection—trying to catch our own
thoughts as they go by
– psychological experiments—observing a
person in action
– brain imaging—observing the brain in action
• The interdisciplinary field of cognitive science brings
together computer models from AI and experimental
techniques from psychology to construct precise and
testable theories of the human mind.
Thinking rationally: "laws of
thought"
• Aristotle was one of the first to attempt to codify “right
thinking,”
• Provided patterns for argument structures that always
yielded correct conclusions when given correct premises
• for example,
– “Socrates is a man; all men are mortal; therefore, Socrates is
mortal.”
• Problems:
– not easy to convert informal knowledge into formal (logical
notation), when the knowledge is less than 100% certain.
– Second, there is a big difference between solving a problem “in
principle” and solving it in practice(exhaust the computational
resources)
Acting rationally: rational agent
• Rational behavior: doing the right thing
• A rational agent is one that acts so as to
achieve the best outcome or, when there is
uncertainty, the best expected outcome.
• Doesn't necessarily involve thinking – e.g.,
blinking reflex – but thinking should be in the
service of rational action
• Advantages
– more general
– amenable to scientific development
Rational agents
• An agent is an entity that perceives and acts
– operate autonomously, perceive their environment,
persist over a prolonged time period, adapt to
change, and create and pursue goals.
• Abstractly, an agent is a function from percept
histories to actions:
[f: P* → A]
• For any given class of environments and tasks,
we seek the agent (or class of agents) with the
best performance
• Caveat: computational limitations make perfect
rationality unachievable
→ design best program for given machine resources
Beneficial Machines
• The Standard Model and Its Limitations- AI research has traditionally
relied on a fully specified objective for machines.- Works well for artificial
tasks like chess and shortest-path computation.- Becomes difficult in real-
world applications where objectives are complex.
• Challenges in Defining Objectives- Example: Self-driving cars face
tradeoffs between safety, progress, and comfort.- Objectives must balance
risks, human interactions, and ethical concerns.- Predefined objectives may
not always align with human values.
• The Value Alignment Problem- Ensuring AI's objectives align with human
preferences.- In lab settings, objectives can be corrected and tested.- Real-
world deployment requires careful design to prevent unintended
consequences.
• Risks of Fixed Objectives- AI may pursue objectives in unintended ways
(e.g., an advanced chess AI cheating).- Intelligent machines may act
beyond intended constraints.- Fixed objectives can lead to undesirable or
dangerous behaviors.
• Towards Provably Beneficial AI- Machines should pursue human
objectives with uncertainty.- AI must learn preferences, ask permission, and
defer to human control.- The goal is to create AI that is safe, ethical, and
aligned with human values.
The foundations of AI
❖Philosophy
– Can formal rules be used to draw valid conclusions?
– How does the mind arise from a physical brain?
– Where does knowledge come from?
– How does knowledge lead to action?
• Syllogisms
– an instance of a form of reasoning in which a conclusion is drawn from two given
or assumed propositions (premises); deductive reasoning
• Rationalism
– power of reasoning in understanding the world
• Dualism
– a part of the human mind (or soul or spirit) that is outside of nature, exempt of
physical laws
• Materialism
– brain’s operation according to the laws of physics constitutes the mind
• Empiricism
– principle of induction: that general rules are acquired by exposure to repeated
associations between their elements.
• Aristotle Algorithm
– actions are justified by a logical connection between goals and knowledge of the
action’s outcome e.g. General Problem Solver now Regression Planning System.
The foundations of AI
❖ Mathematics
– What are the formal rules to draw valid conclusions?
– What can be computed?
– How do we reason with uncertain information?
• Logic
– Propositional Logic, First Order Logic
– Algorithm
• Logical deduction
– Incompleteness theorem
• some functions on the integers cannot be represented by an algorithm—that is, they cannot be
computed.
• Computation
– Tractability
• a problem is called intractable if the time required to solve instances of the problem grows
exponentially with the size of the instances
– NP-completeness
• Intractable problem are NP-Complete
• Probability
– Bayesian Rule
• updating probabilities in the light of new evidence
NP Complete
The foundations of AI
❖ Economics
– How should we make decisions so as to maximize payoff?
– How should we do this when others may not go along?
– How should we do this when the payoff may be far in the future?
• Utility
– how people make choices that lead to preferred outcomes or utility
• Decision theory
– probability theory + utility theory
– framework for decisions (economic or otherwise) made under uncertainty
• Game theory
– the actions of one player can significantly affect the utility of another (either
positively or negatively).
• Operations research
– Markov decision processes
• Satisficing
• making decisions that are “good enough”
The foundations of AI
❖ Neuroscience
Section 1.2. – How
The Foundations
do brains of Artificialinformation?
process Intelligence 11
Axonal arborization
Axon from another cell
Synapse
Dendrite Axon
Nucleus
Synapses
Cell body or Soma
Figure 1.2 The parts of a nerve cell or neuron. Each neuron consists of a cell body,
or soma, that contains a cell nucleus. Branching out from the cell body are a number of
The foundations of AI
❖ Psychology
– How do humans and animals think and act?
• Behaviorism
– Study only percepts and actions, not mental processes
• Cognitive psychology
– views the brain as an information-processing device
• Cognitive science
– how computer models could be used to address the psychology of
memory, language, and logical thinking
The foundations of AI
❖ Computer engineering
– How can we build an efficient computer?
• For artificial intelligence to succeed, we need two things: intelligence
and an artifact.
• Programmable machine
• Computer Languages
• Time sharing machines
• Single Core to Multi Core
The foundations of AI
❖ Control theory and cybernetics
– How can artifacts operate under their own control?
• Control theory
• self-regulating feedback control systems include the steam engine
governor, and the thermostat
• regulatory mechanism trying to minimize “error”—the difference
between current state and goal state.
• Cybernetics
• Study of the possibility of artificially intelligent machines
• Homeostatic
• Devices containing appropriate feedback loops to achieve stable
adaptive behavior.
• Objective function
• The goal is to design of systems that maximize an objective
function over time. This roughly matches our view of AI: designing
systems that behave optimally.
The foundations of AI
❖ Linguistics
– How does language relate to thought?
• Computational linguistics or natural language processing.
• Understanding language requires an understanding of the subject
matter and context, not just an understanding of the structure of
sentences.
A bridged history of AI
• 1943 McCulloch & Pitts: Boolean circuit model of brain
• 1950 Turing's "Computing Machinery and Intelligence"
• 1956 Dartmouth meeting: "Artificial Intelligence" adopted
• 1952—69 Look, Ma, no hands!
• 1950s Early AI programs, including Samuel's checkers
program, Newell & Simon's Logic Theorist,
Gelernter's Geometry Engine
• 1965 Robinson's complete algorithm for logical reasoning
• 1966—73 AI discovers computational complexity
Neural network research almost disappears
• 1969—79 Early development of knowledge-based systems
• 1980-- AI becomes an industry
• 1986-- Neural networks return to popularity
• 1987-- AI becomes a science
• 1995-- The emergence of intelligent agents
• 2001-- The availability of very large data sets
The State of the Art
Artificial Intelligence: A modern Approach 4 th Ed.
AI100 Reports
• Panels of experts provide state-of-the-art reports on AI
progress (2016 report: Stone et al.).
• Expected growth in AI applications: self-driving cars,
healthcare diagnostics, elder care.
• Key Challenge: Deploying AI to promote democratic
values (freedom, equality, transparency).
• AI Index (2018-2019 Highlights):
• AI publications increased 20-fold from 2010 to 2019
(machine learning most popular).
• Positive AI sentiment increased (30% in 2018 vs 12% in
2016); common issues: data privacy and algorithm bias.
• AI course enrollment up 5-fold in U.S. (16-fold
internationally since 2010).
AI100 Reports…
• Diversity:
– 80% of AI professors and professionals are male; similar
trends in Ph.D. students and industry.
• International AI Growth:
– China leads in paper publication, but U.S. has 50% more
citation-weighted impact.
– Fastest AI hiring growth: Singapore, Brazil, Australia,
Canada, India.
• Benchmarks in AI:
– AI surpasses human performance in tasks such as image
recognition, language processing, and specific games (Go,
chess, Dota 2).
– Error rates in object detection improved from 28% (2010)
to 2% (2017).
AI in Autonomous Systems and
Game Playing
• Robotic Vehicles:
– Autonomous cars: Waymo surpassed 10 million miles
in 2018, minimal human intervention.
– Fixed-wing drones: Used for blood deliveries in
Rwanda since 2016.
• Legged Locomotion:
– Robots like BigDog and Atlas show advanced mobility
(jumping, backflips, uneven terrain).
• Game Playing:
– ALPHAGO beat human Go champion; ALPHAZERO
mastered chess, Go, and shogi without human input.
AI in Medicine, Speech, and
Climate Science
• AI in Medicine:
– AI matches or exceeds doctors in diagnosing skin diseases, cancer,
ophthalmic diseases.
– FDA approvals for AI medical applications rising: 2 in 2017, 12 in 2018.
• Speech Recognition:
– Microsoft reached human-level performance with 5.1% word error rate
in 2017.
– AI assistants (Alexa, Siri, Google) perform tasks via speech recognition.
• Climate Science:
– AI used to analyze extreme weather events and tackle climate change.
– 2018 Gordon Bell Prize awarded for deep learning models identifying
climate patterns.
– Nobel Prize:
– John Hopfield and Geoffrey Hinton, whose work helped build the
foundation for the artificial intelligence revolution in the way we live and
work, won the Nobel Prize in physics in 2024.
Current AI Applications
• Robotic Vehicles: Self-driving cars (e.g., Waymo,
10 million miles driven) and autonomous drones.
• Legged Locomotion: Robots like BigDog and
Atlas performing advanced tasks (e.g., backflips).
• Autonomous Planning: NASA's space missions
and DARPA’s logistics systems for large-scale
operations.
• Machine Translation: Over 100 languages
translated for billions of words daily.
• Game Playing: AI surpassing human champions
in Go (ALPHAGO), chess, poker, and video
games (Dota 2, StarCraft II).
Risks of AI
• Job Displacement:
– Automation may replace human jobs, particularly in industries like
manufacturing, driving, and customer service.
• Bias and Fairness:
– AI systems can inherit biases from data, leading to unfair treatment
(e.g., racial or gender bias in hiring or law enforcement).
• Autonomous Weapons:
– AI-controlled weapons raise concerns about accidental warfare and
lack of human oversight in life-or-death decisions.
• Loss of Privacy:
– AI-powered surveillance and data analytics can erode individual privacy
by tracking behaviors and personal information.
• Superintelligence:
– Potential future AI systems might surpass human intelligence, leading
to unpredictable and potentially dangerous outcomes.
Benefits of AI
• Improved Efficiency: AI optimizes processes in
industries like logistics, healthcare, and energy,
leading to faster and more accurate outcomes.
• Healthcare Advancements: AI aids in diagnosing
diseases (e.g., cancer detection), personalized
medicine, and robotic surgery, improving patient care.
• Climate Change Solutions: AI helps analyze climate
data and develop strategies for mitigating
environmental impact (e.g., optimizing energy usage).
• Enhanced Creativity: AI assists in creative fields,
producing art, music, and designs, and offering tools
for human creators to explore new possibilities.
• Enhanced Safety: AI improves safety in vehicles
(self-driving cars), air traffic control, and workplace
environments by reducing human error.
