(বিষয়বিবিক অংশ)

ICT
 Mentor

Md Ferdos Kabir

General Education(ICT)
38th BCS
 Today’s Topic:

Lecture: 6

Artificial Intelligence
Overview
▪ Introduction to AI
▪ Knowledge Representation
▪ Propositional and First-Order Logic
▪ Search Techniques in AI
▪ Probabilistic Reasoning
▪ Natural Language Processing (NLP)
▪ Expert Systems
▪ Introduction to Machine Learning
▪ Practical Applications of Machine
Learning
What is AI?

Science of making computers think and act

like humans
Examples:
Google Maps → shortest route
Playing chess against computer
ChatGPT answering questions

AI = Computer + Human-like Intelligence

Why AI is Important?

Helps humans in decision making Saves

time and effort
Works in complex areas:
• Disease detection
• Self-driving cars
Old AI vs. New AI
Old AI (Symbolic/Rule-based):
Fixed rules: “IF condition THEN action”
Example: MYCIN (Medical Expert System)

New AI (Machine Learning/Deep Learning)

Learns from data
Examples:
• Facebook face recognition
• Siri / Alexa voice assistant

Old AI = Rules | New AI = Data

MCQ
Q1. What does Artificial Intelligence mainly try to do?
a) Make computers faster
b) Make computers think and act like humans
c) Make computers store large data
d) None of the above

Q2. Which of the following is an example of *new AI*?

a) Rule-based expert system
b) Facebook face recognition
c) IF-THEN statement system
d) None of the above
Knowledge Representation

The way information/knowledge is stored in

a computer
Goal: Represent real world in a form that
computer can use

Example:
Human: “All birds can fly. A sparrow is a bird.
→ Sparrow can fly.”
AI stores this as rules/facts
Why Knowledge Representation is
Important?
AI needs facts + reasoning
Without KR, AI is just a program
Helps in decision making and problem
solving
Types of Knowledge

1. Declarative Knowledge → 'What is true'(facts)

Example: Paris is the capital of France

2. Procedural Knowledge → 'How to do

something' (steps)
Example: How to drive a car
Approaches to Knowledge
Representation
1. Semantic Networks → Graph of nodes &
edges
2. Frames → Data structures with slots (e.g.,
Car frame)
3. Rules → IF condition THEN action
4. Ontologies → Defines relationships
between concepts
MCQ
Q1. Which of the following is Declarative
Knowledge?
a) How to cook rice
b) Paris is the capital of France
c) How to ride a bicycle
d) Steps of solving a math problem

Q2. Which method uses graph-like nodes & edges?

a) Frames
b) Semantic Networks
c) Rules
d) Ontologies
What is Logic in AI?

Logic = a way of reasoning

AI uses logic to make decisions and deduce
new facts
Propositional Logic
Deals with simple statements (true/false)
Examples:
• "It is raining." true or false
• "The sky is blue." true
Connectives: AND (∧), OR (∨), NOT (¬), IMPLIES
(→)

Example:
P: It is raining
Q: The ground is wet
Rule: P → Q
First-Order Logic (FOL)
Extends propositional logic with objects,
relations, quantifiers
Predicates: properties (e.g., Human(x))
Quantifiers:
• Universal (∀): "For all"
• Existential (∃): "There exists"

Example:
∀x (Human(x) → Mortal(x))
Human(Socrates) → Mortal(Socrates)
Why Logic is Important in AI?

Helps AI reason like humans

Used in expert systems, NLP, knowledge
bases
MCQs
Q1. Which connective is represented by (¬)?
a) AND
b) OR
c) NOT
d) IMPLIES

Q2. Which quantifier means "For all"?

a) ∀
b) ∃
c) →
d) ¬
What is Search in AI?

AI often needs to find solutions by exploring

possibilities
Examples:
• Solving a maze
• Google Maps finding shortest path
Problem Formulation

Initial state → Where to start

Goal state → What we want to achieve
Operators → Actions that move from one
state to another
Path cost → Measure of efficiency (distance,
time, etc.)
Types of Search – Uninformed

No extra knowledge, just explores

possibilities
Examples:
• Breadth-First Search (BFS): explores level
by level
• Depth-First Search (DFS): goes deep first,
then backtracks
• Uniform Cost Search: chooses least-cost
path
Types of Search – Informed

Uses extra knowledge (heuristics) to guide

search
Examples:
• Greedy Search: chooses path that looks
closest to goal
• A* Search: uses both path cost + heuristic
Why Search is Important in AI?

Many AI problems are search problems

(games, planning, navigation)
Efficient search = faster and smarter AI
MCQ
Q1. Which search explores level by level?
a) DFS
b) BFS
c) Greedy Search
d) A* Search

Q2. Which search algorithm combines path cost +

heuristic?
a) BFS
b) DFS
c) A*
d) Uniform Cost Search
Probabilistic Reasoning

AI often faces uncertainty

Probabilistic reasoning helps make decisions
with incomplete info

Example:
Person coughs → Possible flu (70% chance)
AI uses probability to guess the most likely
cause
Basics of Probability in AI

Probability (P) = likelihood of event

Range: 0 ≤ P(Event) ≤ 1
• 0 = Impossible
• 1 = Certain

Example:
Tossing coin → P(Heads) = 0.5, P(Tails) = 0.5
Bayes' Theorem

Formula:
P(A|B) = [P(B|A) * P(A)] / P(B)

Meaning: Probability of A given B

Example:
Disease probability after test result
Helps update belief with new evidence
Bayesian Networks

Graphical models showing variables &

dependencies

Example:
Weather affects Sprinkler & Rain
Rain + Sprinkler affect Wet Grass
Why Probabilistic Reasoning?

Handles real-life uncertainty Used in:

• Medical diagnosis
• Spam filters
• Speech recognition
• Self-driving cars
MCQ
Q1. Probability values always lie between:
a) -1 and 1
b) 0 and 1
c) 1 and 100
d) None of the above

Q2. Which tool is used in AI to reason under

uncertainty?
a) DFS
b) Bayes' Theorem
c) Semantic Networks
d) Frames
What is NLP?

Teaching computers to understand and

process human language
Bridge between human language and
computer language

Examples:
Google Translate
Siri / Alexa voice assistants
Chatbots
Why is NLP Important?

Enables computers to communicate directly

with humans
Used in:
• Search engines
• Voice assistants
• Customer support
• Healthcare
Main Tasks in NLP
Syntax Analysis (Parsing): Grammar &
structure
Example: "The cat sat on the mat."

Semantics Analysis: Meaning of

words/sentences
Example: Bank = river bank vs. money bank

Pragmatics & Context: Meaning in context

Example: "I’m feeling blue" = sad
Applications of NLP

Machine Translation (Google Translate)

Sentiment Analysis (Positive/Negative)
Chatbots (Customer support)
Speech Recognition (Siri, Alexa)
MCQ
Q1. Which of the following is an application of NLP?
a) Google Translate
b) DFS
c) Sorting algorithms
d) Page replacement

Q2. In NLP, which task deals with meaning of sentences?

a) Syntax Analysis
b) Semantic Analysis
c) Sorting
d) Pathfinding
What is an Expert System?

A computer program that mimics human

expert decision-making
Works using:
• Knowledge base (facts + rules)
• Inference engine (reasoning)

Examples:
MYCIN (medical diagnosis)
DENDRAL (chemical analysis)
Components of Expert Systems

1. Knowledge Base: Stores facts & rules

2. Inference Engine: Applies rules to facts for
conclusions
3. User Interface: Communicates with user
Advantages of Expert Systems

Consistency in decisions
Works 24/7
Useful in medicine, engineering,
troubleshooting
Limitations of Expert Systems

Cannot learn by themselves

Limited to stored knowledge
Expensive to build & maintain
MCQ
Q1. Which is a famous medical expert system?
a) DENDRAL
b) MYCIN
c) Siri
d) Google Translate

Q2. Which component applies rules to facts?

a) Knowledge Base
b) User Interface
c) Inference Engine
d) Database
What is Machine Learning?

A branch of AI where computers learn from

data
Goal: Systems improve automatically with
experience

Examples:
Email spam filter
Netflix recommendations
AI vs ML

AI: Broad field → Making machines

intelligent
ML: Subfield of AI → Machines learn from
data

All ML is AI, but not all AI is ML

Types of Machine Learning –
Supervised
Learns from labeled data (input + output)
Example: Predicting house prices from
size/location
Types of Machine Learning –
Unsupervised
Learns from unlabeled data (no correct
answers)
Example: Grouping customers into clusters
Types of Machine Learning –
Reinforcement
Learns by trial and error using rewards
Example: Training robots, AlphaGo game AI
Applications of ML

Healthcare: Disease prediction

Finance: Fraud detection
Education: Personalized learning
Self-driving cars
Recommendation systems
MCQ
Q1. Which best describes ML?
a) Writing fixed rules
b) Learning from data and improving
c) Using logic only
d) None of the above

Q2. Which ML type learns from labeled data?

a) Supervised Learning
b) Unsupervised Learning
c) Reinforcement Learning
d) None of the above
Why Applications of ML Matter?

ML is powerful because it solves real-world

problems
Helps automate tasks, make predictions,
improve decisions
Applications in Healthcare

Disease prediction (e.g., cancer detection)

Personalized medicine
Applications in Finance

Fraud detection in banking

Stock market predictions
Applications in Education

Personalized learning tools

Student performance prediction
Applications in Transportation

Self-driving cars (Tesla, Waymo)

Traffic prediction (Google Maps)
Applications in Entertainment

Netflix / YouTube recommendations

Music and movie suggestions
Applications in Security

Face recognition systems

Spam filtering in emails
Why Important for Students

ML is used in almost every industry

Opens career opportunities in Data Science,
AI, Robotics, etc.
MCQ
Q1. Which of the following is an application of ML in
finance?
a) Spam filtering
b) Fraud detection
c) Grammar checking
d) Movie recommendation

Q2. Self-driving cars are an example of ML in:

a) Security
b) Entertainment
c) Transportation
d) Education
References
Rich & Knight – Artificial Intelligence (McGraw-Hill)
Standard reference, with conceptual summaries useful for MCQ prep.

R.B. Patel – Artificial Intelligence and Soft Computing

Contains direct MCQs with answers for practice.
Thank You