Introduction to Artificial Intelligence: Concepts and Applications

Ms. Rashmi Bhat
Asst. Professor
St. John College of Engineering
and Management
ARTIFICIAL
INTELLIGENCE

What is Artificial Intelligence?
1
History of Artificial Intelligence
2
Intelligent Systems
3
Components of AI Program
4
Foundations of AI
5
Contents
6 Sub-areas of AI
Applications of AI
7
8 Current Trends in AI

John McCarthy (1955):
"AI is the science and engineering of
making intelligent machines."
What is Artificial
Intelligence?
Marvin Minsky (1965):
"AI is the science of making machines
do things that would require
intelligence if done by humans."

AI is the study of intelligent agents—systems
that perceive their environment and take
actions to maximize their chances of success.
What is Artificial
Intelligence?
"AI is the ability of a computer program to
learn and make decisions"

To Create Expert Systems − The systems which exhibit
intelligent behavior, learn, demonstrate, explain, and advice
its users.
goals of Ai
To Implement Human Intelligence in Machines − Creating
systems that understand, think, learn, and behave like
humans.

examples of Artificial Intelligence
Media streaming
Chatbots
Smart assistants
E-Payments
Search algorithms
Social media feeds
Smart cars
Navigation apps
Facial recognition
Text editors

Thinking
Humanly
Mimicking human
thought processes
(e.g., cognitive
models).
Thinking
Rationally
Using formal
reasoning and
logic (e.g.,
automated
theorem proving).
Acting
Rationally
Taking actions to
achieve goals
efficiently (e.g.,
reinforcement
learning systems).
FOUR approaches of ai
Acting
Humanly
Performing tasks
indistinguishably
from a human
(e.g., passing the
Turing Test).

Definition:
Building systems that perform tasks in a way that is indistinguishable from human behavior.
Turing Test approach proposed by Alan Turing to evaluate if a
machine can exhibit behavior indistinguishable from a human.
“Can a machine think?”
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.
Acting humanly

Acting humanly
Challenges:
Difficulty in accurately modeling human cognition.
The computer would need the following capabilities:
Natural Language Processing to communicate successfully in a human language;
Knowledge Representation to store what it knows or hears;
Automated Reasoning to answer questions and to draw new conclusions;
Machine Learning to adapt to new circumstances and to detect and extrapolate patterns.
A total Turing test, which requires interaction with objects and people in the real world.
To pass the total Turing test, a robot will need
Computer Vision and speech recognition to perceive the world;
Robotics to manipulate objects and move about.

Example:
ChatGPT: Capable of conversational interaction that mimics human responses.
Siri and Alexa: AI systems designed to provide human-like assistance.
Applications:
Customer service chatbots, virtual assistants.
Acting humanly

Definition:
Mimicking human thought processes to understand and replicate how humans think.
Thinking humanly
The cognitive modeling approach
Human thought process can be learnt in three ways:
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.

Challenges:
Difficulty in accurately modeling human cognition.
Thinking humanly
Example:
Cognitive Simulation Models:
Predict how people solve puzzles by
mimicking thought patterns.
ELIZA:
An early chatbot that attempted to
simulate a human therapist's responses.

Definition:
Using formal reasoning to achieve goals, based on the "laws of thought."
The “laws of thought” approach
Logic-based AI focuses on deriving conclusions from a set of premises.
The laws of thought were supposed to govern the operation of the mind; their study
initiated the field called logic.
Probability allows rigorous reasoning with uncertain information, but does not generate
intelligent behavior.
Thinking Rationaly

Example:
Automated Theorem Provers:
Systems like Prolog that solve logical problems.
Medical Diagnosis Systems:
Use logical rules to infer diseases based on symptoms.
Challenges:
Real-world problems often require handling uncertainty, which traditional logic struggles
to address.
Thinking Rationaly

Definition:
Designing systems that act to achieve the best possible outcomes, given available information.
Acting Rationaly
The Rational Agent Approach
An agent is just something that acts.
A 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
A rational agent is one that acts so as to achieve the best outcome or, when there is
uncertainty, the best expected outcome.
Rational agents aim to maximize utility or minimize costs in a given situation.

Advantages of Rational Agent Approach over the other approaches
It is more general than the “laws of thought” approach
It is more amenable to scientific development.
Example:
AlphaGo: Combines deep learning and
reinforcement learning to play Go efficiently.
Self-driving Cars: Use rational decision-
making to navigate traffic and avoid
collisions.
Applications:
Autonomous vehicles, strategic game-playing bots, and smart assistants.
Acting Rationaly

1943:
Warren McCulloch and Walter Pitts propose the first
mathematical model for a neural network.
1950:
Alan Turing publishes "Computing Machinery and Intelligence,"
proposing the Turing Test to evaluate machine intelligence.
Early Foundation HIstory
of
AI

1956:
The term "Artificial Intelligence" is coined during the
Dartmouth Conference by John McCarthy.
Key attendees include Marvin Minsky, Herbert Simon, and Allen
Newell.
Focus: General problem-solving and symbolic reasoning.
Birth of AI HIstory
of
AI

1957:
Frank Rosenblatt designed the Perceptron, an early neural network
model.
1961:
The first industrial robot, Unimate, was installed in a General
Motors factory.
1966:
ELIZA, an early natural language processing program, was created
by Joseph Weizenbaum.
1969:
Marvin Minsky and Seymour Papert publish "Perceptrons,"
highlighting limitations of early neural networks, leading to the AI
Winter.
Early Achievements and Challenges HIstory
of
AI

1972:
The Prolog programming language is developed for logic
programming.
1976:
MYCIN, an expert system for diagnosing bacterial infections,
demonstrates AI in healthcare.
1979:
Stanford Cart, one of the first autonomous robots, navigates a room
avoiding obstacles.
The Era of Expert Systems HIstory
of
AI

1994:
Development of intelligent agents becomes a key research focus.
1997:
IBM's Deep Blue defeats Garry Kasparov, marking the first AI victory
against a reigning world chess champion.
Practical Applications and Milestones
HIstory
of
AI
2006:
Geoffrey Hinton popularizes deep learning, leading to
advancements in neural networks.
2009:
Google begins testing self-driving cars.
The Rise of Machine Learning

2011:
IBM Watson wins the quiz show Jeopardy, showcasing NLP and
knowledge retrieval.
2012:
AlexNet, a deep convolutional neural network, wins the ImageNet
competition, revolutionizing computer vision.
2014:
Chatbot Eugene Goostman claims to pass the Turing Test.
AI Renaissance HIstory
of
AI

2016:
AlphaGo defeats Go champion Lee Sedol, a milestone in
reinforcement learning.
2020:
OpenAI launches GPT-3, a groundbreaking language model.
2021:
DeepMind’s AlphaFold solves the 50-year-old protein folding
problem.
2023:
GPT-4 and similar models demonstrate advanced capabilities in
NLP, reasoning, and multimodal tasks.
The Age of Deep Learning HIstory
of
AI

Intelligent
Systems
Introduction to Intelligent Systems
Definition:
Intelligent systems are computer-based systems that perceive,
reason, learn, and act intelligently in their environment to
achieve specific goals
Key Characteristics:
Adaptive and responsive.
Operate autonomously.
Decision-making capabilities.
Examples:
Virtual assistants (e.g., Alexa, Siri).
Autonomous robots.

Intelligent
Systems
Types of Intelligent Systems
Expert Systems:
Specialized in a domain
to solve problems.
e.g., MYCIN for medical
diagnosis.
Autonomous Systems:
Act independently in real
time
e.g., drones, robotic
vacuum cleaners.
Interactive Systems:
Engage users
e.g., voice assistants.
Hybrid Systems:
Combine multiple
capabilities
e.g., smart assistants in
IoT.

Intelligent Systems
Healthcare: Diagnosis
and personalized
medicine.
Transportation:
Autonomous vehicles
and traffic management.
Education: Adaptive
learning platforms.
Manufacturing: Robots in
assembly lines.
Ethics: Ensuring fairness
and avoiding bias.
Security: Protecting
intelligent systems from
cyber threats.
Scalability: Managing
increasing complexity in
environments.
Applications Challenges

Intelligent
Systems
Components of Intelligent Systems
1. Sensors and Perception:
Collect data from the environment (e.g., cameras, microphones).
Example: Self-driving cars use LIDAR and cameras to perceive
surroundings.
2. Knowledge Representation:
Organizes and stores data for reasoning (e.g., semantic
networks, ontologies).
3. Decision-Making and Reasoning:
Derive conclusions and make choices.
Example: AI in healthcare diagnosing diseases.
4. Learning Capabilities:
Use machine learning to improve performance over time.
Example: Chatbots learning from user queries.

Intelligent
Systems
Categorization of Intelligent Systems
Based on
Capabilities
Based on
Functionalities
Narrow or Weak AI
General or Strong AI
Super AI
Reactive Machines
Limited Memory
Theory of Mind
Self Awareness

Intelligent
Systems
Based on
Capabilities
Narrow or Weak AI
Definition:
It is a type of AI which is able to perform a dedicated task with
intelligence.
Characteristics:
Limited to predefined functions.
Cannot perform tasks outside their domain.
Examples:
Virtual assistants like Siri and Alexa.
Recommendation systems (e.g., Netflix, Amazon).
Image recognition tools.

Intelligent
Systems
Based on
Capabilities
General or Strong AI
Definition:
AI systems with the ability to understand, learn, and perform
any intellectual task that a human can do.
Characteristics:
Mimics human intelligence and reasoning.
Still largely theoretical and under development.
Examples:
Hypothetical AI systems capable of autonomous reasoning
and problem-solving.
Fictional depictions: Jarvis (Iron Man).

Intelligent
Systems
Based on
Capabilities
Super AI
Definition:
An advanced form of AI that surpasses human intelligence in
all aspects, including creativity, decision-making, and
emotional intelligence.
Characteristics:
Self-aware and capable of autonomous reasoning.
Hypothetical and not yet achieved.
Examples:
Conceptual AI systems in science fiction (e.g., Skynet from
The Terminator).

Intelligent
Systems
Key Differences Between Narrow AI, General AI, and Super AI
Based on
Capabilities

Intelligent
Systems
Based on
Functionalities
Reactive Machines
Definition:
Systems that respond to environmental stimuli without storing
past information.
Characteristics:
No memory or ability to learn.
Operate in real-time.
Examples:
IBM’s Deep Blue: A chess-playing AI that evaluates moves in
real-time without learning from past games.

Intelligent
Systems
Based on
Functionalities
Limited Memory
Definition:
Systems that can use past experiences to inform current
decisions for a limited period.
Characteristics:
Learning capabilities with temporary memory.
Widely used in applications today.
Examples:
Self-driving cars: Use data like traffic signals, road conditions,
and nearby vehicles to make decisions.

Intelligent
Systems
Based on
Functionalities
Theory of Mind
Definition:
Systems capable of understanding human emotions, beliefs,
and intentions.
Characteristics:
Focus on human interaction.
Still in experimental stages.
Examples:
Advanced robots interacting with humans empathetically
(e.g., AI companions).

Intelligent
Systems
Based on
Functionalities
Self-Aware Machines
Definition:
AI systems with self-awareness, consciousness, and an
understanding of their own existence.
Characteristics:
Fully autonomous reasoning.
Hypothetical and not yet realized.
Examples:
“Sonny" from the movie I, Robot (2004)
Ultron from Avengers: Age of Ultron (2015)

Learning
Reasoning and
Decision Making
Problem Solving
Processing
Language (NLP)
Perception

Learning: The Heart of AI
Enables AI systems to improve and adapt over time.
Types of Learning:
Supervised Learning:
Training with labeled data.
Example: Predicting house prices based on historical data.
Unsupervised Learning:
Identifying patterns in unlabeled data.
Example: Customer segmentation in marketing.
Reinforcement Learning:
Learning through trial and error with feedback.
Example: Robots learning to walk or AI mastering games like Chess.

Reasoning and Decision Making: How AI & Thinks Acts
AI systems analyze data and make informed decisions.
Key Features:
Draws conclusions from pre-programmed rules or learned models.
Uses probabilistic methods for uncertain scenarios.
Example:
Financial AI systems deciding whether to approve loans.
Diagnostic tools recommending treatments based on symptoms.

Problem Solving: Solving Complex Problems with AI
The ability of AI to identify problems and find solutions.
Strategies:
Algorithms: Step-by-step methods to solve problems.
Heuristics: Problem-solving shortcuts.
Applications:
Gaming (e.g., AI playing chess or Go).
Technical troubleshooting in IT systems.

Perception: How AI Interacts with the World
AI collects and interprets data using sensors and processing algorithms.
Capabilities:
Vision: Recognizing objects in images or videos.
Speech: Understanding spoken language.
Hearing: Interpreting audio signals
Example:
Autonomous cars detecting traffic signals and pedestrians.
Voice-activated assistants understanding commands.

Processing Language: NLP in AI
Enables AI to understand, generate, and respond to human language.
Capabilities:
Text analysis and comprehension.
Speech recognition and synthesis.
Applications:
Chatbots responding to customer queries.
Virtual assistants like Siri or Alexa.

Foundations of AI
Philosophy
Mathematics
Economics
Neuroscience
Psychology
Computer
Engineering
Control Theory
And Cybernetics
Linguistic
1
1
1
3
4
5
6
7
8

Philosophy
Philosophy answers the following questions:
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?
Philosphy is the study of fundamental nature of knowledge i.e. truth, reality and
existence which are considered to solve specific problems.
Solving Specific problem is a basic thing in Artificial Intelligence.

In AI, Mathematics and Statistics is
important for:
Proving the theorems
Writing algorithms
Computations
Decidability
Modelling uncertainty
Learning from data
Agent Programming in AI requires knowledge of Formal Logic and Probability for
planning and learning.
Also, Computation is required for analyzing relation and implementation.
For writing actions for agents, knowldge in formal representation is the major
requirement.
Mathematics and Statistics
These subjects answer:
What are the formal rules to draw the
valid conclusions?
What can be computed?
How do we reason with uncertain
information?

Economics
To develop AI products, one should make decisions for:
Economics deals with
Investing the amount of money,
Maximization of utility with minimal investment.
When to invest? How to invest? How much to invest? Where to invest?
To answer these questions, one should have knowledge about
Decision Theory
Game Theory
Operation Research etc.

The brain consists of nerve cells or Neurons.
Neurons are responsible for:
Individual’s thoughts, his actions and consciousness of his brain.
Neuroscience is the study of nervous system, particulary the human brain
Neuroscience
Human brain is different and more powerful compared to other creatures.
While developing AI system, Use of neuroscience answers the question
How do brains process the information?

Neuroscience
Fig: Parts of Neurons and their functions

Psychology / cognitive science
Psychology is a scientific method to study the human vision
Problem solving skills Behavior of people Perception
Conginitve information processing
Psycology answers following:
How do human and animals thinnk and act?
Knowledge representation

Computer Science
Logic and inference theory
Algorithms
Proramming languages
Software system building
Computer Science and Engineering
Amount of computing power to train top machine learning algorithms and utilization
has been doubled every 100 days
Computer Hardware changed for AI applications:
Graphics Processing Unit (GPU)
Tensor Processing Unit (TPU)
Wafer Scale Engine (WSE)
Super Computers and Quantum Computers are able to solve complicated AI problems.
Software:
Operating System, Programming languages and tools to write modern programs.

AI has founded many ideas in Modern Computer Science, including
Time sharing machines,
Interactive interpreters
High performance PCs
Rapid Development Environments
Linked List data types
automatic Storage management
Key concepts of symbolic, functional, declarative and object oriented programming
Computer Science and Engineering
Computer Science answers:
How can we build fast and efficient computers?

Speech Recognition enables the machines to understand spoken language and
translate it into machine readable format.
linguistics
Linguistic answers:
How does language relate to thoughts?
Computers perform specific tasks on the commands given in spoken language.
It includes:
Speech to Text
Text to Speech

control theory
Examples of control theory :
Self-controlling machines
Self-regulating feedaback control machines
Submarines
Control theory helps the system to analyze, define, debug and fix errors by itself.
Knowledge representation, grammar and NLP are important to develop AI
applications.
The tool of logical inference and computation provide the language, vision and
symbolic planning of agent programming.
Control theory answers following:
How can artifcats operate on their own control ?

Sub-areas of ai
Algorithms that allow systems to learn from data and make predictions or decisions
without explicit programming.
Key Techniques:
Supervised learning, unsupervised learning, reinforcement learning, and deep learning.
Enables computers to understand, interpret, and respond to human languages.
Key Areas:
Text analysis, speech recognition, machine translation, sentiment analysis..
Machine Learning (ML)
Natural LanguageProcessing (NLP)

Sub-areas of ai
Rule-based systems that emulate human decision-making in specific domains.
Key Components:
Knowledge base, inference engine.
Enables a machine to understand the spoken language and translate into a machine-
readable format.
Key Areas:
System control or navigation system, Industrial application, Voice dialing system etc.
Expert Systems
Speech Recognition (NLP)

Sub-areas of ai
Allows machines to interpret and make decisions based on visual data (images or
videos).
Key Techniques:
Object detection, image recognition, and video analysis.
Designing intelligent robots capable of performing tasks in physical environments.
Key Areas:
Path planning, manipulation, human-robot interaction.
Computer Vision
Robotics

Applications of AI
Applications:
Diagnosis of diseases using AI-powered imaging tools.
Personalized treatment recommendations.
Virtual health assistants for patient support.
Drug discovery and development.
Predictive analytics for patient outcomes..
Examples:
IBM Watson, Google DeepMind for Healthcare.
AI in Healthcare

Applications of AI
Applications:
Fraud detection using pattern recognition.
Credit scoring and risk assessment.
Automated trading systems.
Personalized banking assistance (chatbots).
Predictive analytics for market trends.
Examples:
PayPal’s fraud prevention system, JPMorgan’s COiN platform.
AI in Finance

Applications of AI
Applications:
Personalized product recommendations.
Inventory management and demand forecasting.
Virtual shopping assistants.
Dynamic pricing models.
AI-powered customer service.
Examples:
Amazon's Alexa, recommendation systems like those on
Netflix.
AI in Retail and E-Commerce

Applications of AI
Applications:
Autonomous vehicles and self-driving cars.
Traffic management and optimization.
Predictive maintenance of vehicles.
Route planning and navigation.
Examples:
Tesla Autopilot, Google Maps.
AI in Transportation

Applications of AI
Applications:
Intelligent tutoring systems (ITS).
Personalized learning experiences.
Automated grading and feedback.
AI tools for curriculum design.
Examples:
Duolingo, Carnegie Learning
AI in Education

Applications of AI
Applications:
Content recommendation systems.
AI-generated music, art, and stories.
Video and image enhancement.
Real-time language translation for media content.
Examples:
Spotify recommendations, Netflix’s AI-driven personalization.
AI in Entertainment and Media

Applications of AI
Applications:
Threat detection and prevention.
AI-based user authentication systems.
Anomaly detection in network traffic.
Examples:
Darktrace, Symantec’s AI tools.
AI in Cybersecurity

Applications of AI
Applications:
AI-driven opponents for interactive gaming.
Procedural content generation.
Realistic NPC (non-player character) behavior.
Examples:
DeepMind’s AlphaGo, AI in games like The Sims.
AI in Gaming

Introduction to Artificial Intelligence: Concepts and Applications

Introduction to Artificial Intelligence: Concepts and Applications

More Related Content

What's hot (20)

Similar to Introduction to Artificial Intelligence: Concepts and Applications (20)

More from Rashmi Bhat (18)

Recently uploaded (20)

Introduction to Artificial Intelligence: Concepts and Applications