21CSC206T-Artificial Intelligence

UNIT – 1

Introduction to Artificial Intelligence

02/03/2025 Dr. P. Elumalaivasan 1

 Introduction to Artificial Intelligence

• Introduction
1. Artificial Intelligence – History and Foundation
2. Birth of Artificial Intelligence
3. AI Techniques
4. Problem Solving with AI
5. AI Models
6. Data Acquisition and Learning Aspects in AI

02/03/2025 Dr. P. Elumalaivasan 2

 Introduction
Purpose:
• AI is to empower machine to solve the problems by making the machine
intelligent. It referred with reference in relevance with human intelligence.
Definition:
• Rich and Knight – AI is the study of how to make computers do things which, at
the moment, people do better.
• George Boole – Scientist, have le the foundation of AI logic years ago.
• A broad categorization of AI functions and objectives includes reasonable
decision-making, demonstration of intelligence like humans and computational
modelling to solve decision problems.

02/03/2025 Dr. P. Elumalaivasan 3

 Contd…
• w.r.t broad categorisation, AI can be defined in different ways as follows:
1. Machines which can think and have a capability to react like human beings
2. Systems that respond intelligently in the same way as the humans do
3. Computational models to solve various complex decision-making problems
4. Study of intelligent agents
• The horizon of AI includes techniques for knowledge transmission, knowledge
representation, automated reasoning, and this is used to empower machines to behave
intelligently.
• AI is actually an ensemble of technologies, interactions and allied platforms which takes
part in helping machine to demonstrate intelligence and reasonability.
• The human way of thinking, responding and decision-making is the only expectations and
reference in this process.

02/03/2025 Dr. P. Elumalaivasan 4

 History and Foundation
• Since last 60 years, research has been mainstream about AI. But AI philosophy is as old as
one thousand years (Statistics, use of formal systems).
• Aristotle tried to describe and represent human process of thinking and decision-making
using symbols.
• 20th century – idea in fiction started to realise in the form of computer.
• Indian and Greek Philosophers developed various methods for formal reasoning. This was a
structured approach towards problem solving.
• In 190s, Zuse devised high-level programming language and wrote the first chess program
that could demonstrate chess playing.
• Allan Turing devised a simple test of intelligence in 1950, where the response of machine is
expected to be intelligent enough so that it is difficult to find out whether it is machine
sitting on the other side.
• In 1956, John McCarthy insisted and made AI as topic for conference at Dartmouth.
•02/03/2025
IN 1958, he (MIT) invented Lisp language popular for AI-related programming.
Dr. P. Elumalaivasan 5
 Birth of Artificial Intelligence
• Joh McCarthy at Dartmouth college worked on research in the areas of automata theory,
neural nets and study of intelligence.
• McCarthy gave a new term for AI, i.e., ‘computational rationality’.
• Dartmouth is a place where formal workshop and conference on AI took place.
• Then the words like intelligence, knowledge, reasoning, thoughts, cognition and learning
started to formulate a platform for AI research.
• AI is one of the major components of intelligent systems.
• Example - Water tap, washing machines with fuzzy logic, traffic management.

02/03/2025 Dr. P. Elumalaivasan 6

 AI Techniques
• AI deals with a large spectrum of problems. The spectrum of AI applications is spread
across the domains, and even across the complexities of problems.
• This includes the following:
1. Various day-to-day practical problems
2. Different identification and authentication problems with their applications in
security
3. Various classification problems resulting in decision-making
4. Interdependent and cross-domain problems

02/03/2025 Dr. P. Elumalaivasan 7

• AI techniques need to built from the problem-solving perspectives. The points raising need
for AI technique are discussed below:
1. Need for analysis of voluminous and large amount of data. This data may not be
confined to a single domain but may spread across the domains.
2. The analysis should be followed by the characterisation of miscellaneous data, then
mapping of this data with reference to built-in knowledge, and then building the
knowledge further in this process.
3. Dealing with the constantly changing scenarios and situations, and the dynamic nature
of data, the system and technique should react to the new scenario and situation. The
situations are dynamic in nature, and static handling may not be useful.
4. The way in which data appears, the way it is used, the way it is organised and the way it
should be used are different. Blindly using the data as it comes may result in wrong
decisions.
5. Though in some cases, the huge data is available, but the relevant data is limited.
Identification of relevant data, irrelevant data and outliers, and further, effective
knowledge building based on limited relevant data are challenges in front of AI
techniques.
02/03/2025 Dr. P. Elumalaivasan 8
• The main objective of AI techniques is to capture knowledge based on the data and
information.

• There are different scenarios and the relevant knowledge based on the data information.
There are different scenarios and the relevant data is captured. The AI techniques need to
handle different problems.

• The broad categorisation of these problems can be as follows:

1. Structured problems

2. Unstructured problems

3. Linear problems

4. Non-linear problems
02/03/2025 Dr. P. Elumalaivasan 9
 Problem Solving with AI
• AI has been very well used to solve structured problems. The well-structured problems
are some of the very commonly faced problems during day-to-day life. These problems
yield a right answer or right inference when an appropriate algorithm is applied.
• While ill-structured problems are the problems which do not yield a particular answer. In
this case, there is possibility of more than one answer, and even a particular situation
decides the correctness of the answer.
• Interestingly, ill-structured problems represent many of the real-world problems.

02/03/2025 Dr. P. Elumalaivasan 10

 Well structured Problems
• Some of the well-structured problems are given below:
1. Solving a quadratic equation to find the value of X
2. Calculating path of the trajectory when a missile is fired
3. Calculating speed of ball when it reaches to batsman

Ill-structured Problems

• Some of the Ill-structured problems are given below:

1. Predicting how to dispose wet waste safely
2. Analysis of theoretical prepositions and adequacy of the same in a particular
scenario
3. Identifying the security threats in big social gatherings

02/03/2025 Dr. P. Elumalaivasan 11

Behaviour of Problem

Ill-structured problem analysis

Well-structured problem analysis

02/03/2025 Dr. P. Elumalaivasan 12
 AI Models

Chess: A complex scenario

Model building and complexity

02/03/2025 Dr. P. Elumalaivasan 13

Semiotic models
The basic components of a semiotic model include:
1.Signs: These are the basic units of meaning. A sign consists of two parts - the signifier
(the form of the sign, such as a word or an image) and the signified (the concept or
meaning associated with the sign).
2.Semiotic Code: A set of conventions or rules that govern the use and interpretation of
signs within a particular context. Different cultures, disciplines, or systems may have their
own semiotic codes.
3.Encoding and Decoding: The process of creating and interpreting signs. Encoding is the
creation of signs, while decoding is the interpretation of signs. Miscommunication can
occur if there is a mismatch between the encoding and decoding processes.
4.Syntagmatic and Paradigmatic Relations: Syntagmatic relations involve the combination
of signs in a sequence, such as words in a sentence. Paradigmatic relations involve the
selection of one sign over another within a particular category, such as choosing a
specific word or symbol to represent a concept.
02/03/2025 Dr. P. Elumalaivasan 14
Statistical models
1.Variables: In a statistical model, there are variables, which are characteristics or
quantities that can be measured or observed.
2.Parameters: Parameters are the unknown constants in a statistical model that need to be
estimated from the data. They represent the relationships between variables.
3.Probability Distribution: Statistical models often assume a probability distribution for the
data. This distribution describes the likelihood of different outcomes or observations.
4.Assumptions: Statistical models are built based on certain assumptions about the
underlying data and the relationships between variables.

02/03/2025 Dr. P. Elumalaivasan 15

5. Regression Models: One common type of statistical model is the regression model,
where the relationship between a dependent variable and one or more independent
variables is explored.
6. Hypothesis Testing: Statistical models are often used for hypothesis testing, where
researchers assess whether observed data are consistent with a particular hypothesis or if
there are significant differences or relationships between variables.
7. Model Evaluation: Statistical models need to be evaluated for their goodness of fit and
predictive performance. Machine Learning Models: In recent years, machine learning
models, which are a subset of statistical models, have gained popularity. These models use
algorithms to learn patterns from data without explicit programming.

02/03/2025 Dr. P. Elumalaivasan 16

 Data acquisition and learning
aspects in AI
 Knowledge discovery- Data Mining and Machine Learning

 Computational Learning Theory (COLT)

 Neural and evolutionary computation

 Intelligent agent and multi-agent systems

02/03/2025 Dr. P. Elumalaivasan 17

 Knowledge discovery - Data Mining and
Machine Learning
• Data mining is actually the core step in Knowledge Discovery in
Databases (KDD) process.
• Some preprocessing steps before data mining and post processing
steps after data mining are to be completed to transform the raw
data as useful knowledge.
• KDD is an iterative process that transforms raw data into useful
information.

02/03/2025 Dr. P. Elumalaivasan 18

1. Data cleaning

• Noise and irrelevant data are removed from the large data set.
• This is a very important preprocessing step because the outcome would be dependent on
the quality of selected data.
Data cleaning is
• Removal of duplicate records
• Enter logically correct values for missing record
• Remove unnecessary data fields
• Standardize data format
• Update data in a timely manner etc.

02/03/2025 Dr. P. Elumalaivasan 19

2. Data transformation
• With the help of dimensionality reduction or transformation methods, the number of
effective variables is reduced and only useful features are selected to depict data more
efficiently based on the goal of the task.
• In short, data is transformed into appropriate form making it ready for data mining step.
3. Selection of data mining task
• Based on the objective of data mining, appropriate task is selected.
• Some common data mining tasks are classification, clustering, association rule discovery,
sequential pattern discovery, regression and deviation detection.
• We can choose any of these tasks based on whether we need to predict information or
describe information.
02/03/2025 Dr. P. Elumalaivasan 20
4. Selection of data mining algorithm
• Data mining is the actual search for patterns from the data available using the selected
data mining method.
• Appropriate method(s) is to be selected for looking for patterns from the data.
• The model and parameters that might be appropriate for the method are to be decided.
• Some popular data mining methods are decision trees and rules, relational learning
models, example based methods etc.

02/03/2025 Dr. P. Elumalaivasan 21

5. Pattern evaluation
• This is a post processing step in KDD which interprets mined patterns and relationships.
• If the pattern evaluated is not useful, then the process might again start from any of the
previous steps, thus making KDD an iterative process.
6. Knowledge consolidation
• This is the final step in Knowledge Discovery in Databases (KDD).
• The knowledge discovered is consolidated and represented to the user in a simple and
easy to understand format.
• Mostly, visualization techniques are being used to make users understand and interpret
information.

02/03/2025 Dr. P. Elumalaivasan 22

• The mining process includes data cleaning, preprocessing, identifying, interpreting the
patterns, understanding the application and generating the target data
• It acts as a tool and holds core part in business Intelligence
Machine learning
• Machine learning is a field concerned with study of algorithms that will improve its
performance with experience, main focus is on improving performance of agent.

02/03/2025 Dr. P. Elumalaivasan 23

Computational Learning Theory (COLT)
• By defining formal mathematical models, the efficiency and complexity
in terms of computation, prediction and feasibility of algorithms are
analyzed.
• Computation learning theory finds its importance in machine learning,
pattern recognition and statistics.
• There are two frameworks for analysing the patterns
1. Probably Approximately Correct (PAC)
It identifies the class of hypothesis that can/cannot be learnt
2. Mistake bound
It tries to learn target functions to series of trails.

02/03/2025 Dr. P. Elumalaivasan 24

Neural and evolutionary computation
• This method enabled to speed up the mining of data
• Evolutionary computing is related to the study of biological properties
like genetic algorithms.
• Telecom domain to financial decision making, with optimization as base
criteria.
• In case of neural computing, neural behaviour of human brain is
stimulated to enable machine to learn.
• ANN is configured for specific application like pattern recognition or
classification problem.

02/03/2025 Dr. P. Elumalaivasan 25

Intelligent agent and multi-agent systems

• These type of intelligent system allows timely decision making in complex scenarios.
• An agent is simple software program that assists user.
• An intelligent agent is flexible in terms of its action to get the desired output.
• It is goal directed, reacts with environment and acts accordingly.
• Complex tasks and decision-making demand combination of more than one percept
of different intelligent systems which can be done by Multi-Agent System (MAS).
• In MAS every agents capacity and its computation efficiency is exploited so that
overall performance is improved.

02/03/2025 Dr. P. Elumalaivasan 26

Multi perspective integrated intelligence

• For problem solving each and every individual have their own perspective.
Some extra information may/may not present in the different perspective.
• So based on that the decision making can be done.
• Exploiting and utilizing information from different perspectives to build up an
intelligent system giving accurate results or helps in decision making called
Multi Perspective Intelligence (MPI) frame work.
• Taking feedback.
• Collected information can be discrete or continuous in nature.
• This approach works in association with respect to the application it is chosen
for.
• There is a fuzzy line of distinction between them.
02/03/2025 Dr. P. Elumalaivasan 27
 Problem Solving
• Problem Solving:
• Finding answer for some unknown situations.
• Types – simple and complex.
• Simple:
• Can be solved by a deterministic procedure, guarantee of a solution, problem
mapped into an algorithm can be executed by the computer, eg., solving a quadratic
equation.
• Complex:
• Partial data, lack of information, e.g, robot (searching some object or loading boxes
in trucks).
• Human:
• Can reason out, perceive, learn to solve a problem.
• While drawing conclusions, we can use the statistical methods, the mathematical
modelling processes to get best solutions.
• AI:
• Mapping of these intellectual abilities
02/03/2025 into the machine to get the best solutions. 28
Dr. P. Elumalaivasan
Methods:
• General purpose:
• Applicable to a means-ends analysis, the present situation is compared with the goal to
detect the difference. It searches in action space to select an action that will reduce the
difference.
• Eg., if a person wants to cover 50 km distance, the problem solving process will search
from memory for auto-rickshaws, buses, bicycles and so on for the means of transport.
• It would then discard walking and flying to reduce the search space for the next space.
• Special purpose:
• Problem is modeled with various assumptions, which are specific to that problem. Specific
features of the solution are used in this method.
• Eg., classifying legal documents with reference to a particular criminal case.
• In Artificial Intelligence:
• Algorithm using historical knowledge can be treated as expert.
• Eg., during diagnosis of a disease the doctors feed the diagnosis process and the disease-
related symptoms.
• Problem-solving can also assist in planning and decision-making. Planning and decisions are
the key parts of an intelligent system. Dr. P. Elumalaivasan
02/03/2025 29
 Problem-Solving Process
• Problem:
• Used in a situation, when the desired objective is not obvious.
• Initial situation is unknow initially and it consists of sequence of intermediate
objective.
• may vary individual to individual.
• Problem-solving is a process of generating solutions for a given situation.

Problem-solving process
02/03/2025 Dr. P. Elumalaivasan 30
• Problem – encompasses computational tasks also (e.g, matrix inverse, solving quadratic
equations).
• Artificial Intelligence:
• Problem is formulated and solved by searching for a solution in space of possible
solutions (e.g, include NLP, Machine Learning, game palying and so on).
• Problem- may include storage, retrieval, information extraction, decision-making and
so on.
• The term problem can be defined with the following conditions:
• Every problem is defined in a context. In this context, it has certain assumptions
under initial conditions.
• Every problem has a well defined objective.
• Solution to every problem consists of a set of activities. Each activities changes the
state of problem, i.e., from the present state to the new state. This new state is
closer to the solution state. Finally, initial state approaches the goal situation.
• Previous knowledge and domain knowledge both are used as the resources during
different states in the solution process.
02/03/2025 Dr. P. Elumalaivasan 31
• Operation research – the objective is to optimise the operation cost of an item under
boundary conditions.
• Problem-solving techniques involve the following:
• Problem definition
• Problem analysis and representation
• Planning
• Execution
• Evaluating solution
• Consolidating gains

02/03/2025 Dr. P. Elumalaivasan 32

 Formulating Problems

Steps Involved in formulating the Problems

02/03/2025 Dr. P. Elumalaivasan 33

 Problem Identification & Definition

• Frame the questions and answer all to identify the problem.

• E.g. What characterises the class of a problem to be solved?
• What method (or technique) is used in the current stage?
• What information is necessary to specify a particular instance?

02/03/2025 Dr. P. Elumalaivasan 34

 Problem Space

• To be very precise and specific with respect to the problem space

along with the target that we need to achieve.
• Assumptions and limitations of solutions are also specified to
measure the solution quality against the problem objectives and
requirements.

Task Knowledge

• Analysis and representation of task.

02/03/2025 Dr. P. Elumalaivasan 35

 State Space
• General – understand the problem by diagram, description, and so on.
• AI – Use machines to give solutions for a problem by state-space
diagram.
• The problem is defined in terms of State. Each state is the abstraction
of all available information.
• Solution to any problem is the collection of such different states and
set of operations. This collection of states is termed as state space.
• All the states are achieved using the application of actions/operations
to the previous state. During problem solving process, an operator is
applied to a state to move it to the next state.
• Then, another operator is applied and so on till the final state is
achieved. This approach of generating a solutions is called state
space method.
02/03/2025 Dr. P. Elumalaivasan 36
 Example: Consider the problem where three cells in the four-cell
board are filled with single digits and one cell is left blank. The game
is to change positions of the digit and blank cell of the board to arrive
at new board positions. The rule of the game is blank cell can change
the position with a digit by horizontal or vertical movement.

a. Initial State b. Final State

02/03/2025 Dr. P. Elumalaivasan 37

 Problem: To reach from initial state to the final state, with the
minimum number of moves.
Solution: Define the state space, and the operation and action space.
Action space has the operations on the blank cell. Blank cell can move
up (U), down (D), left (L) and right (R). Different board positions
constitute the elements of state space. We designate every state by a
symbol Si. i represents the ith state. S0 is the initial state and Sn is the
final one.

02/03/2025 Dr. P. Elumalaivasan 38

 Algorithm
1. State = initial state; existing state = state
2. While state ǂ final state
i. Existing state = state
ii. Apply operations from set {U, D, L, R} to each state so as to
generate new states.
iii. If new states  existing state
 Existing state = existing state  new states
 State = new state
End while
A Well defined problem, hence, is described in terms of
1. Initial state
2. Goal state
3. List of states
4. Operators or functions that change state or transition of state
5. Path (sequence of states leading to goal state)
6. Path cost (functions that assign a cost to the path)
02/03/2025 Dr. P. Elumalaivasan 39
 Problem Types and Characteristics
Problems can be categorised into different types
• Deterministic or observable state space
• Non-observable
• Non-deterministic or partially observable
• Unknown state space
• Deterministic or observable state space
• Termed as single-state problem
• Goal state is reachable in one single action or sequence of actions
• E.g, vacuum cleaner with sensor
• Non-observable
• Termed as multiple-state problems
• Does not have any information about the state
• Solutions may or may not be reached
• E.g, vacuum cleaner
02/03/2025 Dr. P. Elumalaivasan 40
• Non-deterministic or partially observable
• The effect of action is not clear
• In every new state, some new information is added and then operator acts on the
state
• Solution state is a tree structure of states
• E.g, vaccum clearner with sensor
• Unknown state space
• Problems are typically exploration problems
• State and impact of actions are not known
• There is a need to discover to understand the outcomes of actions
• E.g, online search that involves acting without complete knowledge of the next state or searching
address without map

02/03/2025 Dr. P. Elumalaivasan 41

 Problem Space and Search
• Search is a general algorithm that helps in finding the path in the state space.
• Every problem can be solved with the help of search. It consists of one or more path. The
path may led to the solution or might be dead end.
• In case of dead end, backtrack should occur. The search algorithm, makes use of control
strategy like that or forward or backward search.
• There are two types of strategies:
• Informed search
• Uninformed search

02/03/2025 Dr. P. Elumalaivasan 42

• Informed search:
• Does not guarantee a solution, but there is high probability of getting a solution.
• A heuristic that is specific to the problem is used to control the flow of solution path.
Heuristic is a skill-based technique used to solve the problem.
• It is based on the common sense, rule of thumb, educated guesses or intuitive
judgement.
• It is high order cognitive process. A heuristic function is used in search algorithm to
rank the alternative at each multiple selection path in order to take decision about
which path should be followed.
• When speedy process is required, informed search is preferred over uniformed.
• Uninformed search:
• It does not consider specific nature of the problem
• Very simple and can be generalized to any problem
• Time consuming
• The search methods are evaluated on the basis of completeness, optimality, time
complexity and space complexity.
02/03/2025 Dr. P. Elumalaivasan 43
 Defining the problem as a state-space search
• The problem at hand can be solved by searching the state space to find the path that
leads to a final state from the initial state.
• Solve the problem by checking the steps considering the fact that they might lead us
towards the solution.
• While defining the problem as a state-space search, we mean to say that the goal should
be properly formulated.
• The states and the operators should be clearly indicated.
• There should be transparency in describing the rules and they should be as generalised
as possible.

02/03/2025 Dr. P. Elumalaivasan 44

 Issues in Design of Search Programs
• Apply search, search leads us to goal state.
• Human – can analyse the state space and identify the solution. This happens because of
the cognitive skills that we possess.
• Search space – is traversed. If an error occurs, backtrack should occur so as to evaluate
the alternative path.
• There should be a systematic procedure for forward and backward movement at the
same time systematic method for traversing states in the state space.
• Issues:
• State representation and identifying relationships among states
• Proper selection of forward and backward movement to determine optimal path to
goal state
• Rule selection

02/03/2025 Dr. P. Elumalaivasan 45

 TOY Problems

Tic-Tac-Toe game

02/03/2025 Dr. P. Elumalaivasan 46

 Tic-Tac-Toe
• E = Chances of ‘X’ for winning – Chances of ‘O’ for winning
• E = 0 means Draw
• E = 1 means win

02/03/2025 Dr. P. Elumalaivasan 47

02/03/2025 Dr. P. Elumalaivasan 48
 Travelling Salesman Problem

02/03/2025 Dr. P. Elumalaivasan 49

 Sliding Tile 8-Puzzle Problem

02/03/2025 Dr. P. Elumalaivasan 50