ARTIFICIAL INTELLIGENCE

Dr. Anand M
Assistant Professor
Department of Data Science and Business Systems
SRM Institute of Science and Technology
Introduction to Artificial Intelligence
• Artificial Intelligence is an approach to make a computer, a robot,
or a product to think how smart human think.
• AI is a study of how human brain think, learn, decide and work,
when it tries to solve problems. And finally this study outputs
intelligent software systems.
• The aim of AI is to improve computer functions which are related
to human knowledge, for example, reasoning, learning, and
problem-solving.
• Artificial Intelligence has grown to be very popular in today’s world. It is the
simulation of natural intelligence in machines that are programmed to learn
and mimic the actions of humans. These machines are able to learn with
experience and perform human-like tasks.
Introduction to Artificial Intelligence
• “The science and engineering of making intelligent machines,
especially intelligent computer programs”. -John McCarthy-
• Definitions of Artificial Intelligence according to Authors “Stuart
Russell and Peter Norvig” are vary along two main dimensions.
• Thought processes and reasoning
• Behavior
• A human-centered approach must be in part an empirical science,
involving observations and hypotheses about human behavior.
• A rationalist’s approach involves a combination of mathematics and
engineering. The various groups have both disparaged and helped
each other. Let us look at the four approaches in more detail.
Introduction to Artificial Intelligence
Artificial Intelligence Definition
• An intelligent entity created by humans.
• Capable of performing tasks intelligently without being explicitly
instructed.
• Capable of thinking and acting rationally and humanely.
How do we measure if Artificial Intelligence is
acting like a human?
• Even if we reach that state where an AI can behave as a human does,
how can we be sure it can continue to behave that way? We can base
the human-likeness of an AI entity with the:
• Turing Test
• The Cognitive Modelling Approach
• The Law of Thought Approach
• The Rational Agent Approach
Acting humanly: The Turing Test approach
• The basis of the Turing Test is that the Artificial Intelligence entity
should be able to hold a conversation with a human agent. The
human agent ideally should not able to conclude that they are talking
to an Artificial Intelligence.
• The Turing Test, proposed by Alan Turing (1950), was designed to
provide a satisfactory operational definition of 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.
• This test is used to evaluate a computer acting like humanly.
Acting humanly: The Turing Test approach
Acting humanly: The Turing Test approach
• For current scenarios the computer would need to possess the
following capabilities:
• natural language processing to enable it to communicate successfully in
English
• knowledge representation to store what it knows or hears;
• automated reasoning to use the stored information to answer questions and
to draw new conclusions
• machine learning to adapt to new circumstances and to detect and the
patterns.
Acting humanly: The Turing Test approach
• Total Turing Test includes a video signal so that the interrogator can
test the subject’s perceptual abilities, as well as the opportunity for
the interrogator to pass physical objects ―through the hatch.
• To pass the total Turing Test, the computer will need
• computer vision to perceive objects, and
• Robotics to manipulate objects and move about.
Thinking humanly: The cognitive modeling
approach
• As the name suggests, this approach tries to build an Artificial
Intelligence model based on Human Cognition. To distil the essence of
the human mind, there are 3 approaches:
• Introspection: observing our thoughts, and building a model based on that
• Psychological Experiments: conducting experiments on humans and
observing their behavior
• Brain Imaging: Using MRI to observe how the brain functions in different
scenarios and replicating that through code.
Thinking rationally: The “laws of thought”
approach
• The Laws of Thought are a large list of logical statements that govern
the operation of our mind. The same laws can be codified and applied
to artificial intelligence algorithms.
• The issues with this approach, because solving a problem in principle
(strictly according to the laws of thought) and solving them in practice
can be quite different, requiring contextual nuances to apply.
• Also, there are some actions that we take without being 100% certain
of an outcome that an algorithm might not be able to replicate if
there are too many parameters.
Thinking rationally: The “laws of thought”
approach
Acting rationally: The rational agent approach
• A rational agent acts to achieve the best possible outcome in its
present circumstances.
• According to the Laws of Thought approach, an entity must behave
according to the logical statements. But there are some instances,
where there is no logical right thing to do, with multiple outcomes
involving different outcomes and corresponding compromises. The
rational agent approach tries to make the best possible choice in the
current circumstances. It means that it’s a much more dynamic and
adaptable agent.
• Now that we understand how Artificial Intelligence can be designed
to act like a human, let’s take a look at how these systems are built.
Major Problems Associated with Artificial
Intelligence
• AI is developing with such an incredible speed, sometimes it seems
magical. There is an opinion among researchers and developers that
AI could grow so immensely strong that it would be difficult for
humans to control.
• Humans developed AI systems by introducing into them every
possible intelligence they could, for which the humans themselves
now seem threatened.
Job Loss Problem
• Job loss concerns related to Artificial Intelligence has been a subject of
numerous business cases and academic studies. As per an Oxford Study,
more than 47% of American jobs will be under threat due to automation by
the mid-2030s.
• As per the World Economic Forum, Artificial Intelligence automation will
replace more than 75 million jobs by 2022. Some of the figures are even
more daunting. As per another Mckinsey report, AI-bases robots could
replace 30% of the current global workforce.
• However, these issues related to Job loss and wages can be addressed by
focussing on the following measures.
• Overhauling the education system and giving more focus on skills like Critical
Thinking, Creativity, and Innovation as these skills are hard to replicate.
• Increasing both public and private investment in developing human capital so that
they are better aligned with industry demand.
Safety Problem
• There has always been much furor about safety issues associated with
Artificial Intelligence. When experts like Elon Musk, Stephen Hawking,
Bill Gates among various others express concern related to AI safety
we should pay heed to its safety issues.
• There have been various instances where Artificial Intelligence has
gone wrong when Twitter Chabot started spewing abusive and Pro-
Nazi sentiments and in other instance when Facebook AI bots started
interacting with each other in a language no one else would
understand, ultimately leading to the project being shut down.
Safety Problem
• There are grave concerns about Artificial Intelligence doing something
harmful to humankind. The case in point is autonomous weapons which
can be programmed to kill other humans.
• There are also imminent concerns with AI forming “Mind of their Own” and
doesn’t value human life. If such weapons are deployed, it will be very
difficult to undo its repercussions. The following are the measures that can
be taken to mitigate these concerns.
• We need to have strong regulations especially when it comes to creation or
experimentation of Autonomous weapons
• Global Co-operation on issues concerning such kind of weapons is needed so as to
ensure no one gets involved in the rat race
• Complete transparency in the system where such technologies have experimented is
essential to ensure its safe usage
Trust Related Problem
• As Artificial Intelligence algorithms become more powerful by the day, it
also brings several trust-related issues on its ability to make decisions that
are fair and for the betterment of humankind.
• With AI slowly reaching human-level cognitive abilities the trust issue
becomes all the more significant. There are several applications where AI
operates as a black box.
• Example- in High-Frequency trading even the Program developers don’t
have a good understanding of the basis on which AI executed the trade.
• Some more striking examples include Amazon AI-based algorithm for
same-day delivery which was inadvertently biased against black
neighbourhood.
Trust Related Problem
• Following are few of the measures that can be taken to bridge trust-
related issues in Artificial Intelligence
• All the major Artificial Intelligence providers need to set up guiding rules and
principles related to trust and transparency in AI implementation. These
principles need to be religiously followed by all the stakeholders involved in
Artificial Intelligence development and usage
• All the stakeholders should be aware of the bias which inherently comes with
AI algorithm and should have a robust bias detection mechanism and ways to
handle it
• Awareness is another key factor that plays a major role in bridging the trust
gap. The users should be sensitized about the AI operations, its capabilities
and even the shortfall that is associated with Artificial Intelligence
Computation Problem
• Artificial Intelligence algorithm involves analyzing the humongous amount of data
that require an immense amount of computational power. So far the problem
was dealt with the help of Cloud Computing and Parallel Processing.
• However, as the amount of data increases and more complex deep learning
algorithm comes in the mainstream, the present-day computational power will
not be enough to cater to the complex requirement. We will need more storage
and computational power which can handle crunching exabytes and Zettabytes of
data.
• Quantum Computing can address the processing speed problem in the medium
to long terms
• Quantum computing which is based on concepts of Quantum theory might be the
answer to solving computation power challenges. Quantum computing is 100
Million times faster than a normal computer we use at home. Although currently,
it is in the research and experimental stage. As per an estimate by different
experts, we can see its mainstream implementation in the next 10-15 years.
How Artificial Intelligence (AI) Works?
• Building an AI system is a careful process of reverse-engineering
human traits and capabilities in a machine, and using its
computational prowess to surpass what we are capable of.
• To understand How Artificial Intelligence actually works, one needs to
deep dive into the various sub-domains of Artificial Intelligence and
understand how those domains could be applied to the various fields
of the industry.
• You can also take up an artificial intelligence techniques that will help
you gain a comprehensive understanding.
Artificial intelligence techniques
• Machine Learning : ML teaches a machine how to make inferences and
decisions based on past experience. It identifies patterns, analyses past
data to infer the meaning of these data points to reach a possible
conclusion without having to involve human experience. This automation
to reach conclusions by evaluating data, saves a human time for businesses
and helps them make a better decision.
• Deep Learning : Deep Learning is an ML technique. It teaches a machine to
process inputs through layers in order to classify, infer and predict the
outcome.
• Neural Networks : Neural Networks work on the similar principles as of
Human Neural cells. They are a series of algorithms that captures the
relationship between various underlying variables and processes the data
as a human brain does.
Artificial intelligence techniques
• Natural Language Processing: NLP is a science of reading, understanding,
interpreting a language by a machine. Once a machine understands what
the user intends to communicate, it responds accordingly.
• Computer Vision : Computer vision algorithms tries to understand an
image by breaking down an image and studying different parts of the
objects. This helps the machine classify and learn from a set of images, to
make a better output decision based on previous observations.
• Cognitive Computing : Cognitive computing algorithms try to mimic a
human brain by analyzing text/speech/images/objects in a manner that a
human does and tries to give the desired output.
Tic-Tac-Toe Problem
• Tic-tac-toe is a very popular game for two players, X and O, who take
turns marking the spaces in a 3×3 grid. The player who succeeds in
placing three of their marks in a vertical, horizontal or diagonal row
wins the game.
• In order to solve Tic Tac Toe, we need to go deeper than just to think
about it as a game where two players place X’s and O’s on the board.
Formally speaking, Tic Tac Toe is a zero-sum and perfect information
game. It means that each participant’s gain is equal to the other
participants’ losses and we know everything about the current game
state.
Mathematical Properties
• From a mathematical point of view the game has two very important
properties:
• Property 1: The game admits the player that uses this optimal strategy will
win or draw but it will not lose.
• Property 2: The number of possible different matches is relatively small.
• At the start, the first player can mark any of the 9 spaces. In the following turn the
second player can mark one of the remaining 8 spaces and so on. The game
continues until all the spaces are marked or one of the players win.
• It is then easy to understand that the total number of different matches is lower
than:
• 987....1 = 9! = 362880
• That is a reasonably small number for a computer.
The Algorithm
• From properties 1 and 2 it follows that a practical, and general,
algorithm to win/draw the game is to use the Alpha Beta search.
• At each turn the algorithm evaluates all the possible consequences of
each move (possible due to property 2) and chooses the one that will
ensure a victory or a draw (possible due to property 1).
The Algorithm
• An AI player that chooses each move with the alpha beta search
algorithm will never lose.
• To make the game more realistic it is nice to introduce a stochastic
factor so that each time with a predefined probability the AI player
moves randomly rather than following the alpha beta algorithm.
• This will make the game more realistic as it will make the AI player
more human and sometimes it will lose.
Defining the Problem as a State Space Search

The state space search representation forms the basis of most of the AI method. Its
structure corresponds to the structure problem solving in two important ways:
• If allows for a formal definition of a problem as the need to convert some given
situation into some desired situation using a set of permissible operation.
• It permits us to define the process of solving a particular problem as a combination
of known techniques and search. The general technique of exploring the space to
try to find some path from the current state to a good state.
Defining the Problem as a State Space Search

• Search is very important process in the solution of hard problem for which no
more direct technique are available.
• In order to provide a formal description of a problem it is necessary to do the
following things:
1. Define a state space that contains all the possible configurations of the relevant objects.
2. Specify one or more states within that space that describe possible situation from which the
problem solving process may start. These states are called the initial states.
3. Specify one or more states that would be acceptable as solution to the problem. These states are
called goal states.
4. Specify a set of rules that describe the actions (operators) available.
Production System
• A production system is based on a set of rules about behavior. These
rules are a basic representation found helpful in expert systems,
automated planning, and action selection. It also provides some form
of artificial intelligence.
• Production system or production rule system is a computer program
typically used to provide some form of artificial intelligence, which
consists primarily of a set of rules about behavior but it also includes
the mechanism necessary to follow those rules as the system
responds to states of the world.
Components of Production System
• Global Database: The global database is the central data structure
used by the production system in Artificial Intelligence.
• Set of Production Rules: The production rules operate on the global
database. Each rule usually has a precondition that is either satisfied
or not by the global database. If the precondition is satisfied, the rule
is usually be applied. The application of the rule changes the
database.
• A Control System: The control system then chooses which applicable
rule should be applied and ceases computation when a termination
condition on the database is satisfied. If multiple rules are to fire at
the same time, the control system resolves the conflicts.
Features of Production System in Artificial
Intelligence
• Simplicity: The structure of each sentence in a production system is unique and
uniform as they use the “IF-THEN” structure. This structure provides simplicity in
knowledge representation. This feature of the production system improves the
readability of production rules.
• Modularity: This means the production rule code the knowledge available in
discrete pieces. Information can be treated as a collection of independent facts
which may be added or deleted from the system with essentially no deleterious
side effects.
• Modifiability: This means the facility for modifying rules. It allows the
development of production rules in a skeletal form first and then it is accurate to
suit a specific application.
• Knowledge-intensive: The knowledge base of the production system stores pure
knowledge. This part does not contain any type of control or programming
information. Each production rule is normally written as an English sentence; the
problem of semantics is solved by the very structure of the representation.
Control/Search Strategies
• How would you decide which rule to apply while searching for a
solution for any problem? There are certain requirements for a good
control strategy that you need to keep in mind, such as:
• The first requirement for a good control strategy is that it should cause
motion.
• The second requirement for a good control strategy is that it should be
systematic.
• Finally, it must be efficient in order to find a good answer.
Production System Rules
• Production System rules can be classified as:
• Deductive Inference Rules
• Abductive Inference Rules
• You can represent the knowledge in a production system as a set of
rules along with a control system and database. It can be written as:
• If(Condition) Then (Condition)
• The production rules are also known as condition-action, antecedent-
consequent, pattern-action, situation-response, feedback-result pairs.
Classes of Production System
• Monotonic Production System: It’s a production system in which the application of a rule never
prevents the later application of another rule, that could have also been applied at the time the
first rule was selected.
• Partially Commutative Production System: It’s a type of production system in which the
application of a sequence of rules transforms state X into state Y, then any permutation of those
rules that is allowable also transforms state x into state Y. Theorem proving falls under the
monotonic partially communicative system.
• Non-Monotonic Production Systems: These are useful for solving ignorable problems. These
systems are important from an implementation standpoint because they can be implemented
without the ability to backtrack to previous states when it is discovered that an incorrect path was
followed. This production system increases efficiency since it is not necessary to keep track of the
changes made in the search process.
• Commutative Systems: These are usually useful for problems in which changes occur but can be
reversed and in which the order of operation is not critical. Production systems that are not
usually not partially commutative are useful for many problems in which irreversible changes
occur, such as chemical analysis. When dealing with such systems, the order in which operations
are performed is very important and hence correct decisions must be made at the first attempt
itself.
PROBLEM CHARACTERISTICS
• Heuristics cannot be generalized, as they are domain specific. Production
systems provide ideal techniques for representing such heuristics in the
form of IF-THEN rules.
• Most problems requiring simulation of intelligence use heuristic search
extensively. Some heuristics are used to define the control structure that
guides the search process, as seen in the example described above.
• But heuristics can also been coded in the rules to represent the domain
knowledge. Since most AI problems make use of knowledge and guided
search through the knowledge, AI can be described as the study of
techniques for solving exponentially hard problems in polynomial time by
exploiting knowledge about problem domain.
PROBLEM CHARACTERISTICS
• To use the heuristic search for problem solving, we suggest analysis of
the problem for the following considerations:
• Decomposability of the problem into a set of independent smaller sub
problems
• Possibility of undoing solution steps, if they are found to be unwise
• Predictability of the problem universe
• Possibility of obtaining an obvious solution to a problem without comparison
of all other possible solutions
• Type of the solution: whether it is a state or a path to the goal state
• Role of knowledge in problem solving
• Nature of solution process: with or without interacting with the user
PROBLEM CHARACTERISTICS
• The general classes of engineering problems such as
planning, classification, diagnosis, monitoring and design are
generally knowledge intensive and use a large amount of
heuristics. Depending on the type of problem, the knowledge
representation schemes and control strategies for search are to be
adopted
PROBLEM CHARACTERISTICS
PROBLEM CHARACTERISTICS
• Suppose we are trying to prove a mathematical theorem: first we proceed
considering that proving a lemma will be useful. Later we realize that it is not at
all useful. We start with another one to prove the theorem. Here we simply
ignore the first method.
• Consider the 8-puzzle problem to solve: we make a wrong move and
realize that mistake. But here, the control strategy must keep track of all the
moves, so that we can backtrack to the initial state and start with some new
move.
• Consider the problem of playing chess. Here, once we make a move we never
recover from that step. These problems are illustrated in the three important
classes of problems mentioned below:
• Ignorable, in which solution steps can be ignored.Eg: Theorem Proving
• Recoverable, in which solution steps can be undone.Eg: 8-Puzzle
• Irrecoverable, in which solution steps cannot be undone.Eg: Chess
Issues in the Design of Search Programs
• Each search process can be considered to be a tree traversal. The
object of the search is to find a path from the initial state to a goal
state using a tree.
• The number of nodes generated might be huge; and in practice many
of the nodes would not be needed.
• The secret of a good search routine is to generate only those nodes
that are likely to be useful, rather than having a precise tree.
• The rules are used to represent the tree implicitly and only to create
nodes explicitly if they are actually to be of use.
Issues in the Design of Search Programs
• The following issues arise when searching:
• The tree can be searched forward from the initial node to the goal state or
backwards from the goal state to the initial state.
• To select applicable rules, it is critical to have an efficient procedure for
matching rules against states.
• How to represent each node of the search process? This is the knowledge
representation problem or the frame problem. In games, an array suffices; in
other problems, more complex data structures are needed.
• Finally in terms of data structures, considering the water jug as a
typical problem do we use a graph or tree? The breadth-first structure
does take note of all nodes generated but the depth-first one can be
modified.
Issues in the Design of Search Programs
Issues in the Design of Search Programs
Issues in the Design of Search Programs
Issues in the Design of Search Programs
Problem Reduction Methods
• When a problem can be divided into a set of sub problems, where
each sub problem can be solved separately and a combination of
these will be a solution, AND-OR graphs or AND - OR trees are used
for representing the solution.
• The decomposition of the problem or problem reduction generates
AND arcs. One AND are may point to any number of successor nodes.
All these must be solved so that the arc will rise to many arcs,
indicating several possible solutions. Hence the graph is known as
AND - OR instead of AND.
Problem Reduction algorithm:
 An AI system is composed of an agent and its
environment. The agents act in their
Agents environment. The environment may contain
other agents.
Agent
 • An agent is anything that can perceive its
environment through sensors and acts upon
that environment through effectors.
• A human agent has sensory organs such as
eyes, ears, nose, tongue and skin parallel to
What are the sensors, and other organs such as hands,
Agent and legs, mouth, for effectors.
• A robotic agent replaces cameras and
Environment? infrared range finders for the sensors, and
various motors and actuators for effectors.
• A software agent has encoded bit strings as
its programs and actions.
 • Performance Measure of Agent − It is the
criteria, which determines how successful an
agent is.
• Behavior of Agent − It is the ac on that
agent performs after any given sequence of
percepts.
Agent • Percept − It is agent’s perceptual inputs at a
Terminology given instance.
• Percept Sequence − It is the history of all
that an agent has perceived till date.
• Agent Function − It is a map from the
precept sequence to an action.
 • Rationality is nothing but status of being
reasonable, sensible, and having good sense
of judgment.
Rationality • Rationality is concerned with expected
actions and results depending upon what
the agent has perceived. Performing actions
with the aim of obtaining useful information
is an important part of rationality.
 • An ideal rational agent is the one,
which is capable of doing expected
actions to maximize its
What is Ideal performance measure, on the
Rational basis of −
Agent? • Its percept sequence
• Its built-in knowledge the agent
has.
 • The performance measures, which
determine the degree of success.
Rationality of • Agent’s Percept Sequence till now.
an agent • The agent’s prior knowledge about
the environment.
depends on • The actions that the agent can
the following carry out.
 • Agent’s structure can be viewed as −
The • Agent = Architecture + Agent Program
Structure of • Architecture = the machinery that an agent
executes on.
Intelligent • Agent Program = an implementation of an
Agents agent function.
 • They choose actions only based on the
current percept.
Simple Reflex • They are rational only if a correct decision is
made only on the basis of current precept.
Agents
• Their environment is completely observable.
Condition-Action
Rule − It is a rule
that maps a state
(condition) to an
action
 • They use a model of the world to choose
their actions. They maintain an internal
state.
• Model − knowledge about “how the things
Model Based happen in the world”.
• Internal State − It is a representa on of
Reflex Agents unobserved aspects of current state
depending on percept history.
• Updating the state requires the
informa on about −
• How the world evolves.
• How the agent’s actions affect the world.
Model
Based
Reflex
Agents
 • They choose their actions in order to achieve
goals. Goal-based approach is more flexible
than reflex agent since the knowledge
Goal Based supporting a decision is explicitly modeled,
thereby allowing for modifications.
Agents • Goal − It is the descrip on of desirable
situations.
Goal Based Agents
 • They choose actions based on a preference
(utility) for each state.
• Goals are inadequate when −
Utility Based • There are conflicting goals, out of which only
few can be achieved.
Agents • Goals have some uncertainty of being
achieved and you need to weigh likelihood
of success against the importance of a goal.
Utility Based Agents
 • Discrete / Continuous − If there are a limited
number of distinct, clearly defined, states of the
environment, the environment is discrete (For
example, chess); otherwise it is continuous (For
example, driving).
Properties of • Observable / Partially Observable − If it is possible
to determine the complete state of the environment
Environment at each time point from the percepts it is
observable; otherwise it is only partially observable.
• Static / Dynamic − If the environment does not
change while an agent is acting, then it is static;
otherwise it is dynamic.
 • Single agent / Multiple agents − The environment may contain
other agents which may be of the same or different kind as
that of the agent.
• Accessible / Inaccessible − If the agent’s sensory apparatus
can have access to the complete state of the environment,
then the environment is accessible to that agent.

Properties of • Deterministic / Non-deterministic − If the next state of the

environment is completely determined by the current state
Environment and the actions of the agent, then the environment is
deterministic; otherwise it is non-deterministic.
• Episodic / Non-episodic − In an episodic environment, each
episode consists of the agent perceiving and then acting. The
quality of its action depends just on the episode itself.
Subsequent episodes do not depend on the actions in the
previous episodes. Episodic environments are much simpler
because the agent does not need to think ahead.
 • The games such as 3X3 eight-tile, 4X4
fifteen-tile, and 5X5 twenty four tile puzzles
are single-agent-path-finding challenges.
They consist of a matrix of tiles with a blank
tile. The player is required to arrange the
Single Agent tiles by sliding a tile either vertically or
Pathfinding horizontally into a blank space with the aim
of accomplishing some objective.
Problems • The other examples of single agent
pathfinding problems are Travelling
Salesman Problem, Rubik’s Cube, and
Theorem Proving.