Knowledge Representation
• Knowledge Representation helps to represent
information from the real world for a computer
to understand
• Knowledge representation in AI is not just
about storing data in a database
• Knowledge representation allows a machine to
learn from that knowledge and behave
intelligently like a human being
Need for Knowledge Representation
• One of the primary purposes of Knowledge Representation
is modeling intelligent behavior like humans
• Humans tackle difficult real-world problems by using their
knowledge(from experience, training)
• The problem-solving agents have only limited knowledge
– (eg. actions available, the result of a specific action from a
specific state)
• They don’t know general facts
– (eg. The length of a road cannot be a negative number, two tiles
cannot occupy the same space etc.)
• Intelligent Systems cannot achieve anything useful unless
they contain a large amount of real-world/domain-specific
knowledge
Components of an
Artificial Intelligent System
Components of an
Artificial Intelligent System
1. Perception component retrieves information from
the environment
2. Knowledge Representation component represents
that information
3. Reasoning component utilizes this knowledge to
solve problems
4. Learning Component learns from the captured data
5. Planning components finds a sequence of actions to
achieve the goal state (given an initial state,
preconditions and effects)
6. The Execution component executes the plan
Kinds of Knowledge
• The different kinds of knowledge that need to
be represented in AI include:
1. Objects(Structural Knowledge)
2. Actions(Procedural knowledge)
3. Facts (Declarative Knowledge)
4. Meta-Knowledge
5. Heuristic Knowledge
6. Events (temporal knowledge)
Knowledge Representation
• Formalism: The notation and structure in which
information is expressed (eg. Symbols,
language)
• Knowledge Base: A model represented using
formalisms
– A knowledge base is a collection of facts and beliefs
– Knowledge bases help to model a domain of discourse and
to draw inferences about the objects in the domain
(reasoning)
Desirable Properties of a
Knowledge Representation System
• Representational Adequacy: ability to represent the
required knowledge
• Inferential Efficiency: ability to direct the inferential
mechanisms into the most productive directions (to
respond with limited resources)
• Inferential Adequacy: ability to manipulate the
knowledge represented to produce new knowledge
corresponding to that inferred from the original
• Acquisitional Efficiency: ability to acquire new
knowledge using automatic methods wherever possible
rather than reliance on human intervention
Knowledge Acquisition
Knowledge Engineering
• The process of assessing problems, acquiring
knowledge and building knowledge based
systems
• Activities
1. Problem assessment
2. Data and knowledge acquisition
3. Development of a prototype system
4. Evaluation and revision of the system
5. Development of a complete system
6. Integration and maintenance of the system
Knowledge Representation Formalisms
• Different knowledge representation techniques
include
1. Simple relational knowledge
2. Inheritable Knowledge (Slot and filler structure)
3. Inferential Knowledge (Propositional and
Predicate Logic)
4. Semantic Networks
5. Frames
6. Scripts
7. Conceptual Graphs
Relational Knowledge - Table
• Relational knowledge is made up of objects
consisting of attributes and their corresponding
associated values.
• Example: a table
Relational Knowledge
• Advantages:
– simplest way of storing facts
• Disadvantages:
– little opportunity for inferencing and learning
Inheritable Knowledge
Semantic Network
• The semantic network is a graphical knowledge
representation technique
• A semantic net is a directed graph consisting of
nodes connected by arcs
• The vertices represent concepts and edges
represent semantic relations between the
concepts
Semantic Net Relations
1. Meronymy: A is part of B
2. Holonymy: B has A as a part of itself
3. Hyponymy(troponymy): A is subordinate of B;
A is kind of B
4. Hypernymy: A is superordinate of B
5. Synonymy: A denotes the same as B
6. Antonymy: A denotes the opposite of B
Semantic Nets
• Knowledge Representation in hierarchical structure
has
is a
Scooter Two - wheeler Motor – bike
Brakes Moving – vehicles Engine
Electrical system Fuel - system
is a is a
has
has has
Semantic Nets
animal
skin
fish
swimming
bird
flying
feathers
penguin canary robin
ostrich
walking
Opus
Tweety
yellow
red
white
covered_by
travels_by
isa
isa
isa isa isa isa
covered_by
travels_by
travels_by
travels_by
instance_of
instance_of
colour
colour
colour
Semantic Nets
Semantic Nets
• Advantages:
– Natural way of representation of knowledge,
– simple and easy to understand
• Disadvantages:
– Takes more computational time at runtime,
depend on the creator of the system,
– not intelligent
FRAMES
• Frames are general record like structures which
consist of a collection of slots and slot values and
describe some entity in the world
• The slots may be of any size and type and have
names(attributes) and values or subfields called facets
• Frame are used to represent two types of knowledge
– declarative/factual
– procedural
• A frame that merely contains description about
objects is call a declarative /factual frame
• Frames which have procedural knowledge embedded
in it are called action/procedural frames.
Slots in a Frame
• Slots in a frame can contain
– information for choosing a frame in a situation
– relationship between this and other frames
– procedures to carry out after various slots filled
– default information to use when input is missing
– blank slots
– other frames, which gives a hierarchy
Declarative /Factual frame
Name : Computer Centre
A/c Stationar
y
cupboard
Computer Dumb
terminals
Printer
Name of the frame
Slots in the frame
Action/Procedural Frame
Name : Cleaning the carburetor
Expert (Actor)
Scooter
(Source)
Scooter
(Destination)
Remove
Carburetor
( Task 1)
Clean
Nozzle
( Task 2)
Fix
Carburetor
( Task 3)
who performs
the activity
from where the
action has to begin
where action
has to end
Examples of Frames
Example 1 Example 2
Representing Hierarchy using Frames
Frames
• Advantages:
– Makes the programming easier by grouping the
related data,
– easy to understand and visualize,
– easy to add slots for new attributes and relations,
– easy to include default data and search for missing
values
• Disadvantages:
– cannot be easily processed,
– has a very generalized approach
Scripts
• A script is a structured representation describing
a stereotyped sequence of events in a particular
context
• Scripts are used to organize events in knowledge
bases
Components of Scripts
• Entry conditions that must be true for the script to
be called
• Results or facts that are true once the script has
terminated
• Props or the ”things” that make up the content of
the script
• Roles are the actions that the individual participants
perform
• Scenes which present temporal aspects of the script
Example: Visit to a Restaurant
Example: Visit to a Restaurant
Symbols(Meaning) to describe the Script
Conceptual Dependency Graphs
• Represents knowledge acquired from natural
language input
• Captures the concepts about the events and
represents them in the form of a graph
• A concept may be individual or generic
• The agent and the objects are represented
• The actions are represented using a set of primitive
acts which can be modified by tense
Primitive Acts
Primitive Conceptual Categories
Conceptual Syntax Rules
Conceptual Syntax Rules
Conceptual Syntax Rules
Modifiers
• p -- past
• f -- future
• t -- transition
• ts -- start transition
• tf -- finished transition
• k -- continuing
• ? -- interrogative
• / -- negative
• delta -- timeless
• c -- conditional
• the absence of any modifier implies the present tense
Examples
Examples
Examples

Knowledge Representation Part 1 with answers

  • 1.
    Knowledge Representation • KnowledgeRepresentation helps to represent information from the real world for a computer to understand • Knowledge representation in AI is not just about storing data in a database • Knowledge representation allows a machine to learn from that knowledge and behave intelligently like a human being
  • 2.
    Need for KnowledgeRepresentation • One of the primary purposes of Knowledge Representation is modeling intelligent behavior like humans • Humans tackle difficult real-world problems by using their knowledge(from experience, training) • The problem-solving agents have only limited knowledge – (eg. actions available, the result of a specific action from a specific state) • They don’t know general facts – (eg. The length of a road cannot be a negative number, two tiles cannot occupy the same space etc.) • Intelligent Systems cannot achieve anything useful unless they contain a large amount of real-world/domain-specific knowledge
  • 3.
    Components of an ArtificialIntelligent System
  • 4.
    Components of an ArtificialIntelligent System 1. Perception component retrieves information from the environment 2. Knowledge Representation component represents that information 3. Reasoning component utilizes this knowledge to solve problems 4. Learning Component learns from the captured data 5. Planning components finds a sequence of actions to achieve the goal state (given an initial state, preconditions and effects) 6. The Execution component executes the plan
  • 5.
    Kinds of Knowledge •The different kinds of knowledge that need to be represented in AI include: 1. Objects(Structural Knowledge) 2. Actions(Procedural knowledge) 3. Facts (Declarative Knowledge) 4. Meta-Knowledge 5. Heuristic Knowledge 6. Events (temporal knowledge)
  • 6.
    Knowledge Representation • Formalism:The notation and structure in which information is expressed (eg. Symbols, language) • Knowledge Base: A model represented using formalisms – A knowledge base is a collection of facts and beliefs – Knowledge bases help to model a domain of discourse and to draw inferences about the objects in the domain (reasoning)
  • 7.
    Desirable Properties ofa Knowledge Representation System • Representational Adequacy: ability to represent the required knowledge • Inferential Efficiency: ability to direct the inferential mechanisms into the most productive directions (to respond with limited resources) • Inferential Adequacy: ability to manipulate the knowledge represented to produce new knowledge corresponding to that inferred from the original • Acquisitional Efficiency: ability to acquire new knowledge using automatic methods wherever possible rather than reliance on human intervention
  • 8.
  • 9.
    Knowledge Engineering • Theprocess of assessing problems, acquiring knowledge and building knowledge based systems • Activities 1. Problem assessment 2. Data and knowledge acquisition 3. Development of a prototype system 4. Evaluation and revision of the system 5. Development of a complete system 6. Integration and maintenance of the system
  • 10.
    Knowledge Representation Formalisms •Different knowledge representation techniques include 1. Simple relational knowledge 2. Inheritable Knowledge (Slot and filler structure) 3. Inferential Knowledge (Propositional and Predicate Logic) 4. Semantic Networks 5. Frames 6. Scripts 7. Conceptual Graphs
  • 11.
    Relational Knowledge -Table • Relational knowledge is made up of objects consisting of attributes and their corresponding associated values. • Example: a table
  • 12.
    Relational Knowledge • Advantages: –simplest way of storing facts • Disadvantages: – little opportunity for inferencing and learning
  • 13.
  • 14.
    Semantic Network • Thesemantic network is a graphical knowledge representation technique • A semantic net is a directed graph consisting of nodes connected by arcs • The vertices represent concepts and edges represent semantic relations between the concepts
  • 15.
    Semantic Net Relations 1.Meronymy: A is part of B 2. Holonymy: B has A as a part of itself 3. Hyponymy(troponymy): A is subordinate of B; A is kind of B 4. Hypernymy: A is superordinate of B 5. Synonymy: A denotes the same as B 6. Antonymy: A denotes the opposite of B
  • 16.
    Semantic Nets • KnowledgeRepresentation in hierarchical structure
  • 17.
    has is a Scooter Two- wheeler Motor – bike Brakes Moving – vehicles Engine Electrical system Fuel - system is a is a has has has Semantic Nets
  • 18.
    animal skin fish swimming bird flying feathers penguin canary robin ostrich walking Opus Tweety yellow red white covered_by travels_by isa isa isaisa isa isa covered_by travels_by travels_by travels_by instance_of instance_of colour colour colour Semantic Nets
  • 19.
    Semantic Nets • Advantages: –Natural way of representation of knowledge, – simple and easy to understand • Disadvantages: – Takes more computational time at runtime, depend on the creator of the system, – not intelligent
  • 20.
    FRAMES • Frames aregeneral record like structures which consist of a collection of slots and slot values and describe some entity in the world • The slots may be of any size and type and have names(attributes) and values or subfields called facets • Frame are used to represent two types of knowledge – declarative/factual – procedural • A frame that merely contains description about objects is call a declarative /factual frame • Frames which have procedural knowledge embedded in it are called action/procedural frames.
  • 21.
    Slots in aFrame • Slots in a frame can contain – information for choosing a frame in a situation – relationship between this and other frames – procedures to carry out after various slots filled – default information to use when input is missing – blank slots – other frames, which gives a hierarchy
  • 22.
    Declarative /Factual frame Name: Computer Centre A/c Stationar y cupboard Computer Dumb terminals Printer Name of the frame Slots in the frame
  • 23.
    Action/Procedural Frame Name :Cleaning the carburetor Expert (Actor) Scooter (Source) Scooter (Destination) Remove Carburetor ( Task 1) Clean Nozzle ( Task 2) Fix Carburetor ( Task 3) who performs the activity from where the action has to begin where action has to end
  • 24.
  • 25.
  • 26.
    Frames • Advantages: – Makesthe programming easier by grouping the related data, – easy to understand and visualize, – easy to add slots for new attributes and relations, – easy to include default data and search for missing values • Disadvantages: – cannot be easily processed, – has a very generalized approach
  • 27.
    Scripts • A scriptis a structured representation describing a stereotyped sequence of events in a particular context • Scripts are used to organize events in knowledge bases
  • 28.
    Components of Scripts •Entry conditions that must be true for the script to be called • Results or facts that are true once the script has terminated • Props or the ”things” that make up the content of the script • Roles are the actions that the individual participants perform • Scenes which present temporal aspects of the script
  • 29.
    Example: Visit toa Restaurant
  • 30.
    Example: Visit toa Restaurant
  • 31.
  • 33.
    Conceptual Dependency Graphs •Represents knowledge acquired from natural language input • Captures the concepts about the events and represents them in the form of a graph • A concept may be individual or generic • The agent and the objects are represented • The actions are represented using a set of primitive acts which can be modified by tense
  • 34.
  • 35.
  • 36.
  • 37.
  • 38.
  • 39.
    Modifiers • p --past • f -- future • t -- transition • ts -- start transition • tf -- finished transition • k -- continuing • ? -- interrogative • / -- negative • delta -- timeless • c -- conditional • the absence of any modifier implies the present tense
  • 40.
  • 41.
  • 42.