Knowledge Representation in Artificial Intelligence

Knowledge Representation (KR) plays an essential role in artificial intelligence that enables
systems to organize and interpret information in a way similar to human thinking.
This allows AI systems to process data, make decisions, and tackle problems by keeping
knowledge in a structure form. Just as humans use language or symbols to express thoughts,
AI needs structures to represent the world around it.

Knowledge Representation in AI uses the methods and frameworks to encode and store
knowledge making it accessible to reason and make decisions. It allows machines to process
and use information to understand the world, solve problems, and learn from experiences.

 KR has an impact on simplifying and structuring large amounts of complex data, which makes
it easier to analyze and apply AI.
 Good knowledge representation can boost machine learning by offering improved data
structures that the system can employ to learn and recognize patterns.
 Knowledge representation frameworks can be used in different problem areas, like healthcare,
robotics, or finance letting AI work in various contexts.
 By arranging knowledge well, KR helps make decisions faster and with more information.

Knowledge Representation in AI uses the methods and frameworks to encode and store
knowledge making it accessible to reason and make decisions. It allows machines to process
and use information to understand the world, solve problems, and learn from experiences.

 KR has an impact on simplifying and structuring large amounts of complex data, which makes
it easier to analyze and apply AI.
 Good knowledge representation can boost machine learning by offering improved data
structures that the system can employ to learn and recognize patterns.
 Knowledge representation frameworks can be used in different problem areas, like healthcare,
robotics, or finance letting AI work in various contexts.
 By arranging knowledge well, KR helps make decisions faster and with more information.

Types of Knowledge in AI

Following are the various types of Knowledge −

 Declarative Knowledge: Declarative knowledge refers to facts, statements, or

information that describe "what is known" about a domain. It can be described as static since
the information can be represented as either assertions or truths.
For example, declarative knowledge has facts like "The sky is blue", "Delhi is capital of India", and "A
triangle has three sides". These statements are declarative knowledge because it is a fact that can be
directly expressed and documented.
 Declarative knowledge contains various facts and knowledge regarding the world.
  It answers "what" queries rather than "how" to achieve anything.
 This type of knowledge is easy to convey through statements, databases, or documents.
 Examples include scientific truths, historical events, and general knowledge.
 AI uses declarative knowledge for reasoning, decision-making, and problem-solving.

AI Application: In a question-and-answer system, declarative knowledge is utilized to answer

factual questions such as "What is the capital of France?"

 Procedural Knowledge: Procedural knowledge refers to the knowledge,

of how to perform a task. It involves procedures, methods, or processes that needed
to achieve a task or solve a problem. It focuses on step-by-step procedures rather than
just facts.
For example, solving a quadratic equation is an structured process that includes determining
coefficients, applying the quadratic formula, and simplifying the result.
 Procedural knowledge gives step-by-step instructions for how to do a task.
 It explains "how" to do something rather than just stating the facts.
 This knowledge is acquired through practice and experience.
 Procedural knowledge more difficult to convey explicitly than declarative knowledge.
 It is widely used in artificial intelligence applications like automation, robotics, and
expert systems.
AI Application −
In robotics, procedural knowledge is applied when programming a robot to perform
duties such as assembling auto-mobiles or navigating in a maze.
A robot chef follows rules in a step-by-step manner to prepare a meal. "How To Make
Tea: Boil water. Add tea leaves. Pour into a cup. Add sugar and milk to taste."

Meta Knowledge
Meta-knowledge is a term that refers to "knowledge about knowledge." It allows AI
to understand what it knows, how reliable the information is, and when to apply it. This kind
of knowledge allows AI systems to evaluate and enhance their reasoning and decision-making
abilities.

For instance, if an AI chatbot knows that its answers are sourced from a reliable database, it
will become more confident in its responses.

 Meta-knowledge is the capability of AI to evaluate the correctness and reliability of its own
knowledge.
 It enables AI to determine whether it should use specific rules or facts.
 Crucial in learning, over time it sharpens the artificial intelligence decision to make.
 It is very useful for debugging AI models as it helps detecting gaps or inconsistencies.
 Meta-knowledge improves problem-solving by helping AI in choosing the best reasoning
strategy.
 AI application, a self-driving auto-mobile knows traffic laws and has meta-knowledge to
recognize false sensor data caused by fog. It can then decide whether to slow down or
switch to a backup system.

Heuristic knowledge: Heuristic knowledge is the rule-of-thumb or experience-

based knowledge that helps in problem-solving and decision-making when complete
information is not available. It helps in decision-making when specific rules or formulae are
not available.

Heuristic knowledge often relies on intuition, experience, or common sense.

This knowledge is used in AI to make accurate and quick decisions, particularly in complex or
unclear situations.

Heuristic information reduces computation time by directing the search process toward likely
solutions.

It is widely used in applications such as game AI, medical diagnostics, and optimization
problems where exact answers are hard to compute.

AI Application: In game-playing AI like chess or Go, heuristic knowledge helps the

machine to evaluate board positions and make strategic decisions.

Structural Knowledge: Structural knowledge refers to

the relationships and connections between different concepts or things in a domain. It
helps AI understand how things are related.
 Structural knowledge describes how things are structured and connected.
 This knowledge is often represented as graphs, trees, or networks.
 Structural knowledge is widely used in many applications of artificial intelligence, for example,
in semantic networks, ontologies, and knowledge graphs.
 This type of knowledge assists AI in inferring new relationships from the existing ones which
improves future decision making process quick and accurate.
 Structural knowledge strengthens reasoning and decision-making by presenting information in
an organized manner.

For example, an AI-driven medical diagnosis system understands, A fever is a symptom

of flu. Flu is caused by a viral infection. Antiviral drugs can treat viral infections this way
it connects different entities in a domain.
Cycle of Knowledge Representation in AI

Artificial Intelligent Systems usually consist of various components to display their intelligent
behavior. Some of these components include:

 Perception
 Learning
 Knowledge Representation & Reasoning
 Planning
 Execution

Here is an example to show the different components of the system and how it works:

Example

The above diagram shows the interaction of an AI system with the real world and the
components involved in showing intelligence.

 The Perception component retrieves data or information from the environment. with
the help of this component, you can retrieve data from the environment, find out the
source of noises and check if the AI was damaged by anything. Also, it defines how to
respond when any sense has been detected.
  Then, there is the Learning Component that learns from the captured data by the
perception component. The goal is to build computers that can be taught instead of
programming them. Learning focuses on the process of self-improvement. In order to
learn new things, the system requires knowledge acquisition, inference, acquisition of
heuristics, faster searches, etc.
 The main component in the cycle is Knowledge Representation and Reasoning which
shows the human-like intelligence in the machines. Knowledge representation is all
about understanding intelligence. Instead of trying to understand or build brains from
the bottom up, its goal is to understand and build intelligent behavior from the top-
down and focus on what an agent needs to know in order to behave intelligently. Also,
it defines how automated reasoning procedures can make this knowledge available as
needed.
 The Planning and Execution components depend on the analysis of knowledge
representation and reasoning. Here, planning includes giving an initial state, finding
their preconditions and effects, and a sequence of actions to achieve a state in which a
particular goal holds. Now once the planning is completed, the final stage is the
execution of the entire process.

What to represent in AI: Following are the kind of knowledge which needs to be
represented in AI systems −
 Object Knowledge: Information about physical objects and their properties, for example, "A
laptop has keyboard, mouse, screen". or "A tree has branches, leaves, and roots. "This helps AI
recognize and classify things in its environment.
 Event Knowledge: Knowledge about actions or events, for example, "A traffic light turns red"
or "A user clicks a button." This helps AI understand cause and effect.
 Performance Knowledge: Knowledge of how to do tasks, like "How to Bake a Cake" or "How
to Troubleshoot a Machine." It gives AI the efficiency to accomplish tasks.
 Meta Knowledge: Knowledge of what the system knows, such as understanding when to
update the knowledge base or determining which information is most useful in a given context.
 Factual Knowledge: Verifiable statements or facts, such as "Water boils at 100C." It serves
as the basis of AI reasoning and decision-making.
 Knowledge base: A centralized source of maintaining facts, rules, and procedures, which
enables the AI system to take decisions and address problems based on relevant knowledge.

Applications of Knowledge Representation in AI

Knowledge representation is applied across various domains in AI, enabling systems to

perform tasks that require human-like understanding and reasoning. Some notable applications
include:
1. Expert Systems: These systems use knowledge representation to provide advice or make
decisions in specific domains, such as medical diagnosis or financial planning.
 2. Natural Language Processing (NLP): Knowledge representation is used to understand
and generate human language, enabling applications like chatbots, translation systems, and
sentiment analysis.
3. Robotics: Robots use knowledge representation to navigate, interact with environments,
and perform tasks autonomously.
4. Semantic Web: The Semantic Web relies on ontologies and other knowledge
representation techniques to enable machines to understand and process web content
meaningfully.
5. Cognitive Computing: Systems like IBM's Watson use knowledge representation to
process vast amounts of information, reason about it, and provide insights in fields like
healthcare and research.

Techniques/ Approaches of Knowledge Representation

They are mainly four types of knowledge representation which are as follows –

 Logical Representation
 Semantic Network Representation
 Frame representation
 Production Rules

Logical Representation:

Logical Representation is a fundamental method of communicating knowledge to machines

through a well-defined syntax with precise rules. This syntax should be unambiguous and able
to handle prepositions, making it an ideal way to represent facts. There are two types of logical
representation: Propositional Logic and First-order Logic.

Propositional logic (PL) has basic statements (propositions) that logical operators like AND,
OR, and NOT connect.
First-Order Logic (FOL) expands on PL by adding objects, relationships, and quantifiers such
as (for all) and (exists).

 Propositional Logic, also known as propositional calculus or statement logic, is a formal

system of logic that deals with the relationships between propositions, which are
statements that are either true or false. Propositional logic is based on the Boolean
system, which means that propositions are evaluated as either true or false. In
propositional logic, propositions are combined using logical connectives such as "and",
"or", and "not", and the resulting compound propositions can also be evaluated as true
or false based on the truth values of their component propositions.
 It Represents knowledge as declarative statements (propositions) linked by logical
operators like AND, OR, and NOT. For example, "If it rains (A) AND the ground is
wet (B), THEN the road is slippery (C)." While simple, it struggles with complex
relationships. Often follow the format "IF condition THEN conclusion." For instance,
in a knowledge-based system, you might have:
IF an object is red AND round, THEN the object might be an apple.

 First-order logic (FOL), also known as first-order predicate calculus (FOPC) or first-
order logic with identity, is an extension of propositional logic that allows for the
representation of more complex relationships between objects. In FOL, propositions
are constructed using predicates, which are statements that describe properties or
relations between objects, and quantifiers, which specify the scope of the variables in
the proposition.

FOL allows for more precise and flexible reasoning about the relationships between
objects and is widely used in mathematics, computer science, and philosophy.

In logic, we use symbols and operators to represent concepts like truth, negation, conjunction,
disjunction, implication, quantification, and identity. There are different types of logical
representation like propositional logic, first-order logic, and higher-order logic.

The semantics of logical representation involves assigning meaning to these symbols and
formulas. This is done by defining a set of axioms and rules for manipulating these symbols.

There are several advantages to using logical representation, such as its ability to facilitate
logical reasoning and serve as the foundation for programming languages. However, there are
also some limitations and challenges associated with this method. One disadvantage is that
logical representations can be restrictive and difficult to work with. Additionally, this approach
may not be very intuitive, and the process of inference may not always be efficient.

Example:

 It is Monday.
  The Sun rises from the North (False proposition)
 3+3= 8(False proposition)
 7 is a prime number.

Advantages:

 Logical representation helps to perform logical reasoning.

 This representation is the basis for the programming languages.
 Logical representation enables a clear and precise statement of knowledge, which
removes uncertainty and ambiguity.
 It allows AI systems to infer and reason based on established facts and logical rules.
 This method is mathematically accurate and acceptable for AI implementation in the
forms of expert systems and automated deduction.
 Logical representation is widely utilized in AI to perform theorem proof, knowledge-
based systems, as well as natural language processing.


Disadvantages:

 Logical representations have some restrictions and are challenging to work with.
 This technique may not be very natural, and inference may not be very efficient.
 Logical deduction can be computationally slow and complex, especially in First-Order
Logic, where it requires more computer power.
 This approach does not handle uncertainty well, and it is hard to characterize real-world
scenarios where knowledge is probabilistic or imperfect.
 It is strict and inflexible, demanding exact definitions of all facts and rules, and is hard
to handle flexible or evolving information.
 As the knowledge base grows, the system becomes harder to control, leading to
inefficiencies in reasoning and increased complexity.

Semantic Net Representation: Semantic networks work as an alternative of predicate logic

for knowledge representation. In Semantic networks, you can represent your knowledge in the
form of graphical networks. This network consists of nodes representing objects and arcs which
describe the relationship between those objects. Also, it categorizes the object in different
forms and links those objects.

This representation consist of two types of relations:

 IS-A relation (Inheritance)

 Kind-of-relation

Semantic network is a form of knowledge structuring according to a network of concepts and

the relationships among them. Think of it as a map, where one idea is connected to other
ideas to signify how they're related, for example, "dog" and "animal."
 Semantic networks in AI enable programs to understand and draw conclusions by examining
these connections. For example, if a program understands that "dogs are animals" and "animals
need food," it can deduce that "dogs need food" as well.

For example let us consider a university system in which there are various entities connected
to each other. Let us represent these relations through nodes and arcs.

 John is a student.
 John pursues computer science.
 Computer science is a department.
 All departments belong to the university.
 John owns a laptop.
 John owns a Dell laptop.

This above network allows AI systems to draw new conclusions. For example, Since John is
studying Computer Science, we can deduce that he is connected to the institution via his
department. Since, John has a Dell laptop, its fair to assume he uses it for his academic work.

Advantages of Semantic Network Representation

The key advantages of semantic network representation are as follows −

 Using natural representations of real-world associations makes it easier to store data

and retrieve information.
 It allows for inheritance, meaning AI can deduce characteristics (for instance, if all
birds can fly, then an eagle is capable of flying).
 It's easy to visualize and can be expanded by introducing new concepts and connections.
 Supports reasoning by following connections between related concepts.
 Semantic networks are a natural representation of knowledge.
 Also, it conveys meaning in a transparent manner.
 These networks are simple and easy to understand.


Disadvantages of Semantic Network Representation

The following are the major disadvantages associated with semantic network representation −

 As more concepts and relationships are added, the network grows and becomes increasingly
difficult to manage.
 There is no uniform method for representing relationships, leading to inconsistencies in how
knowledge is illustrated.
 Finding relationships within a large network can require significant time and computational
resources.
 Semantic networks have difficulty representing probabilistic or uncertain knowledge when
compared to probabilistic models.
 Frame Representation:
 Frame representation is a technique for organizing knowledge in a hierarchical structure. A
frame is a structured record that describes an entity in the world by using a collection of
attributes and their corresponding values. In artificial intelligence, frames serve as a data
structure that divides knowledge into substructures by representing typical situations.

Frame representation offers a method to organize details about objects, events, or concepts.
This idea proposes that human memory uses "frames" or "templates" to represent general
situations, objects, or events. Each frame has slots (features) and fillers (instances) that describe
the traits of what it represents.

For example let us consider the definition of "car." The frame for a car might have fields for
color, model, make, and year. These fields could be filled with instances such as red, Model X,
Tesla, and 2023.

Key Components of Frame Representation

The following are the main components of frame representation that help in the successful
structuring and organization of knowledge.

 Frames are structures that hold data representing a specific entity, concept, or situation. They
serve as templates for organizing relevant information.
 Slots are the properties and attributes of a frame. They describe the characteristics of the entity
being represented. For example, in a "person" frame, the slots might include name, age, gender,
and occupation.
 Fillers are the actual values assigned to these slots. In the case of the "person" frame, John Doe
would fill the name slot.
 Frames can be merged to illustrate relationships between entities. For instance, a "car" frame
might be linked with an "owner" frame to indicate ownership.
 Some slots may have default values that are assumed unless explicitly stated differently. For
example, the default value for the color slot in a "car" frame could be black.

Advantages of Frame representation

Frame representation offers several benefits, which are listed below −

 It facilitates programming by organizing relevant information in a comprehensible and

organized form.
 It is flexible and simple to extend, so new features and associations can be introduced with
minimal effort.
 The visualization is straightforward, and people can easily comprehend how the information is
stored and retrievable.
 This technique is suitable for many AI applications, including natural language processing and
machine vision.
 Disadvantages of Frame representation

Despite its usefulness, frame representation has some drawbacks, as mentioned below −

 Finding the relevant information over many frames can be challenging and time-consuming.
 Developing rules to accurately infer new knowledge from frames is difficult, especially for big
systems.
 Representing highly complex or dynamic knowledge with frames can be difficult because it
may necessitate multiple slots and connections.
 Frame-based systems are effective for structured knowledge, but they may not handle
ambiguous or abstract concepts adequately.

Production Rules: Production rules are a set of IF-THEN statements that represent
knowledge. The IF part of a rule is a condition, and the THEN part is an action to be taken if
the condition is met. Production rules can be used to represent a wide range of knowledge,
including facts, procedures, and heuristics.

The production rules system is composed of three key components:

 The set of production rules

 The working memory
 The recognize-act-cycle

The syntax of a production rule consists of IF-THEN statements, and the semantics involve
defining the meaning of the conditions and actions.

Production rules are a knowledge representation technique that consists of a series of if-
then rules that are used to make decisions and solve problems. These rules provide actions
based on conditions and are widely utilized in expert systems and rule-based AI models.

A production rule is written as −

IF (Condition) THEN (Action)

For example, IF temperature > 40C THEN turn on the fan.

 Production rules are used in expert systems to solve problems using rule bases, inference
engines, and reasoning methods such as forward and backward chaining.

Examples of Production Rules

The following are few examples of production rules illustrating how conditions meet certain
actions in AI systems.

 If the patient has a temperature and a cough, then the diagnosis is flu.
 If the words like "win cash" or "free offer", then mark the email as spam.
 If motion is detected at night, then switch on the lights.
 If the user says "hello", then respond with "Hi! "How can I help you?".
 Advantages of Production Rules

The primary advantages of using production rules in AI are as follows −

 Production rules are simple to grasp since they contain IF-THEN expressions.
 Rules can be introduced and withdrawn without affecting the entire system. A smart home
system, for example, can add new rules such as IF the temperature exceeds 30C, then turn on
the air conditioner, without affecting existing rules.
 Used in AI-powered decision-making systems like medical diagnosis and automation.
 Both the rules (knowledge base) and the inference engine (processing) are maintained
independently, making maintenance easier.

Disadvantages of Production Rules

The following are the key disadvantages of production rules in AI −

 As the number of regulations grows, managing them becomes more complex.

 Checking each rule one by one can slow down the system.
 Production rules do not learn from prior experiences unless integrated with machine learning.
 Traditional production rules perform best in clear-cut situations, but they struggle when faced
with uncertain or probabilistic data.