Knowledge Representation Page 1 of 1 What to Represent?

Page 1 of 2

Next: What to Represent? Up: Artificial Intelligence II Previous: Exercises Next: Using Knowledge Up: Knowledge Representation Previous: Knowledge Representation

Knowledge Representation What to Represent?

Let us first consider what kinds of knowledge might need to be represented in AI systems:

Objects
-- Facts about objects in our world domain. e.g. Guitars have strings, trumpets are brass
! What to Represent? instruments.
! Using Knowledge
! Properties for Knowledge Representation Systems Events
! Approaches to Knowledge Representation -- Actions that occur in our world. e.g. Steve Vai played the guitar in Frank Zappa's Band.
" Simple relational knowledge

" Inheritable knowledge Performance

! Inferential Knowledge -- A behavior like playing the guitar involves knowledge about how to do things.
" Procedural Knowledge

! Issue in Knowledge Representation Meta-knowledge

! Summary and the way forward -- knowledge about what we know. e.g. Bobrow's Robot who plan's a trip. It knows that it can
! Further Reading read street signs along the way to find out where it is.

Thus in solving problems in AI we must represent knowledge and there are two entities to deal with:

Facts
[email protected] -- truths about the real world and what we represent. This can be regarded as the knowledge
level
Representation of the facts
which we manipulate. This can be regarded as the symbol level since we usually define the
representation in terms of symbols that can be manipulated by programs.

We can structure these entities at two levels

the knowledge level

-- at which facts are described
the symbol level
-- at which representations of objects are defined in terms of symbols that can be manipulated
in programs (see Fig. 5)

Fig 5 Two Entities in Knowledge Representation

English or natural language is an obvious way of representing and handling facts. Logic enables us to

http://www.cs.cf.ac.uk/Dave/AI2/node32.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node33.html 28/08/01

What to Represent? Page 2 of 2 Using Knowledge Page 1 of 1

consider the following fact: spot is a dog as dog(spot) We could then infer that all dogs have tails
with: : dog(x) hasatail(x) We can then deduce:
Next: Properties for Knowledge Representation Up: Knowledge Representation Previous: What to
hasatail(Spot) Represent?

Using an appropriate backward mapping function the English sentence Spot has a tail can be Using Knowledge
generated.
We have briefly mentioned where knowledge is used in AI systems. Let us consider a little further to
The available functions are not always one to one but rather are many to many which is a what applications and how knowledge may be used.
characteristic of English representations. The sentences All dogs have tails and every dog has a tail
both say that each dog has a tail but the first could say that each dog has more than one tail try Learning
substituting teeth for tails. When an AI program manipulates the internal representation of facts these -- acquiring knowledge. This is more than simply adding new facts to a knowledge base. New
new representations should also be interpretable as new representations of facts. data may have to be classified prior to storage for easy retrieval, etc.. Interaction and inference
with existing facts to avoid redundancy and replication in the knowledge and and also so that
Consider the classic problem of the mutilated chess board. Problem In a normal chess board the facts can be updated.
opposite corner squares have been eliminated. The given task is to cover all the squares on the Retrieval
remaining board by dominoes so that each domino covers two squares. No overlapping of dominoes -- The representation scheme used can have a critical effect on the efficiency of the method.
is allowed, can it be done. Consider three data structures Humans are very good at it.

Many AI methods have tried to model human (see lecture on distributed reasoning)

Reasoning
-- Infer facts from existing data.

If a system on only knows:

! Miles Davis is a Jazz Musician.

! All Jazz Musicians can play their instruments well.
Fig. 3.1 Mutilated Checker
If things like Is Miles Davis a Jazz Musician? or Can Jazz Musicians play their instruments well? are
the first two are illustrated in the diagrams above and the third data structure is the number of black asked then the answer is readily obtained from the data structures and procedures.
squares and the number of white squares. The first diagram loses the colour of the squares and a
solution is not east to see; the second preserves the colours but produces no easier path whereas However a question like Can Miles Davis play his instrument well? requires reasoning.
counting the number of squares of each colour giving black as 32 and the number of white as 30
yields an immediate solution of NO as a domino must be on one white square and one black square, The above are all related. For example, it is fairly obvious that learning and reasoning involve
thus the number of squares must be equal for a positive solution. retrieval etc.

Next: Using Knowledge Up: Knowledge Representation Previous: Knowledge Representation [email protected]
[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node33.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node34.html 28/08/01

Properties for Knowledge Representation Systems Page 1 of 1 Approaches to Knowledge Representation Page 1 of 1

Next: Approaches to Knowledge Representation Up: Knowledge Representation Previous: Using Next: Simple relational knowledge Up: Knowledge Representation Previous: Properties for
Knowledge Knowledge Representation

Properties for Knowledge Representation Systems Approaches to Knowledge Representation

The following properties should be possessed by a knowledge representation system. We briefly survey some representation schemes. We will look at some in more detail in further
lectures. Also some other course (e.g. Expert Systems also deal with related subjects).
Representational Adequacy
-- the ability to represent the required knowledge;
Inferential Adequacy
- the ability to manipulate the knowledge represented to produce new knowledge
corresponding to that inferred from the original;
Inferential Efficiency ! Simple relational knowledge
- the ability to direct the inferential mechanisms into the most productive directions by storing ! Inheritable knowledge
appropriate guides;
Acquisitional Efficiency
- the ability to acquire new knowledge using automatic methods wherever possible rather than
reliance on human intervention.
[email protected]
To date no single system optimises all of the above

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node35.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node36.html 28/08/01

Simple relational knowledge Page 1 of 1 Inheritable knowledge Page 1 of 2

Next: Inheritable knowledge Up: Approaches to Knowledge Representation Previous: Approaches Next: Inferential Knowledge Up: Approaches to Knowledge Representation Previous: Simple
to Knowledge Representation relational knowledge

Simple relational knowledge Inheritable knowledge

The simplest way of storing facts is to use a relational method where each fact about a set of objects Relational knowledge is made up of objects consisting of
is set out systematically in columns. This representation gives little opportunity for inference, but it
can be used as the knowledge basis for inference engines. ! attributes
! corresponding associated values.
! Simple way to store facts.
! Each fact about a set of objects is set out systematically in columns (Fig. 7). We extend the base more by allowing inference mechanisms:
! Little opportunity for inference.
! Knowledge basis for inference engines. ! Property inheritance
" elements inherit values from being members of a class.

" data must be organised into a hierarchy of classes (Fig. 8).

Figure: Simple Relational Knowledge

We can ask things like:

! Who is dead?
! Who plays Jazz/Trumpet etc.?

This sort of representation is popular in database systems.

[email protected]

Fig. 8 Property Inheritance Hierarchy

! Boxed nodes -- objects and values of attributes of objects.

! Values can be objects with attributes and so on.
! Arrows -- point from object to its value.
! This structure is known as a slot and filler structure, semantic network or a collection of
frames.

The algorithm to retrieve a value for an attribute of an instance object:

1. Find the object in the knowledge base

2. If there is a value for the attribute report it
3. Otherwise look for a value of instance if none fail
4. Otherwise go to that node and find a value for the attribute and then report it
5. Otherwise search through using isa until a value is found for the attribute.

http://www.cs.cf.ac.uk/Dave/AI2/node37.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node38.html 28/08/01

Inheritable knowledge Page 2 of 2 Inferential Knowledge Page 1 of 1

Next: Procedural Knowledge Up: Knowledge Representation Previous: Inheritable knowledge

[email protected]
Inferential Knowledge
Represent knowledge as formal logic:

All dogs have tails : dog(x) hasatail(x) Advantages:

! A set of strict rules.

" Can be used to derive more facts.
" Truths of new statements can be verified.
" Guaranteed correctness.

! Many inference procedures available to in implement standard rules of logic.

! Popular in AI systems. e.g Automated theorem proving.

! Procedural Knowledge

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node38.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node39.html 28/08/01

Procedural Knowledge Page 1 of 1 Issue in Knowledge Representation Page 1 of 2

Next: Issue in Knowledge Representation Up: Inferential Knowledge Previous: Inferential Next: Summary and the way Up: Knowledge Representation Previous: Procedural Knowledge
Knowledge
Issue in Knowledge Representation
Procedural Knowledge
Below are listed issues that should be raised when using a knowledge representation technique:
Basic idea:
Important Attributes
! Knowledge encoded in some procedures -- Are there any attributes that occur in many different types of problem?
" small programs that know how to do specific things, how to proceed.

" e.g a parser in a natural language understander has the knowledge that a noun phrase
There are two instance and isa and each is important because each supports property
may contain articles, adjectives and nouns. It is represented by calls to routines that
inheritance.
know how to process articles, adjectives and nouns.
Relationships
Advantages:
-- What about the relationship between the attributes of an object, such as, inverses, existence,
techniques for reasoning about values and single valued attributes. We can consider an
! Heuristic or domain specific knowledge can be represented. example of an inverse in
! Extended logical inferences, such as default reasoning facilitated.
! Side effects of actions may be modelled. Some rules may become false in time. Keeping track band(John Zorn,Naked City)
of this in large systems may be tricky.
This can be treated as John Zorn plays in the band Naked City or John Zorn's band is Naked
Disadvantages: City.
! Completeness -- not all cases may be represented. Another representation is band = Naked City
! Consistency -- not all deductions may be correct.
band-members = John Zorn, Bill Frissell, Fred Frith, Joey Barron,
e.g If we know that Fred is a bird we might deduce that Fred can fly. Later we might discover
that Fred is an emu.
Granularity
-- At what level should the knowledge be represented and what are the primitives. Choosing
! Modularity is sacrificed. Changes in knowledge base might have far-reaching effects. the Granularity of Representation Primitives are fundamental concepts such as holding, seeing,
! Cumbersome control information. playing and as English is a very rich language with over half a million words it is clear we will
find difficulty in deciding upon which words to choose as our primitives in a series of
situations.

If Tom feeds a dog then it could become:

[email protected]
feeds(tom, dog)

If Tom gives the dog a bone like:

gives(tom, dog,bone) Are these the same?

In any sense does giving an object food constitute feeding?

If give(x, food) feed(x) then we are making progress.

But we need to add certain inferential rules.

In the famous program on relationships Louise is Bill's cousin How do we represent this? louise =
daughter (brother or sister (father or mother( bill))) Suppose it is Chris then we do not know if it is
Chris as a male or female and then son applies as well.

http://www.cs.cf.ac.uk/Dave/AI2/node40.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node41.html 28/08/01

Issue in Knowledge Representation Page 2 of 2 Summary and the way forward Page 1 of 1

Clearly the separate levels of understanding require different levels of primitives and these need
many rules to link together apparently similar primitives.
Next: Further Reading Up: Knowledge Representation Previous: Issue in Knowledge
Obviously there is a potential storage problem and the underlying question must be what level of Representation
comprehension is needed.
Summary and the way forward
In this lecture we have hopefully seen the need for knowledge in reasoning programs.
Next: Summary and the way Up: Knowledge Representation Previous: Procedural Knowledge
Many issues need to be considered when deciding on the representation scheme for knowledge.
[email protected]
We introduced the concept of a slot and filler data structure for inheritable knowledge. In the next
two lectures we will be looking at different types of slot and filler representations starting with
semantic nets and frames and moving onto stronger conceptual dependency based representations.

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node41.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node42.html 28/08/01

Further Reading Page 1 of 1 Logic Knowledge Representation Page 1 of 1

Next: Logic Knowledge Representation Up: Knowledge Representation Previous: Summary and the Next: Predicate logic Up: Artificial Intelligence II Previous: Further Reading
way

Further Reading Logic Knowledge Representation

Artificial Intelligence by Rich and Knight covers most topics well. We briefly mentioned how logic can be used to represent simple facts in the last lecture. Here we
will highlight major principles involved in knowledge representation. In particular predicate logic
will be met in other knowledge representation schemes and reasoning methods.
One book dedicated to this field is Knowledge Representation: An AI perspective by H. Reichgelt,
Ablex, New York.
A more comprehensive treatment is given in the third year Expert Systems course. Symbols used
The following standard logic symbols we use in this course are:
The Handbook of Artificial Intelligence (Vol. 1, Ch 3) provides a good concise introduction to most
topics
For all
Bobrow and Collins deal with fundamental issues and also their robot planner in Representation and
Understanding, Academic Press (1975). There exists

Implies

Not

[email protected] Or

And

Let us now look at an example of how predicate logic is used to represent knowledge. There are
other ways but this form is popular.

! Predicate logic
" An example

" Isa and instance relationships

! Applications and extensions
! Further reading
! Exercises

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node43.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node44.html 28/08/01

Predicate logic Page 1 of 1 An example Page 1 of 1

Next: An example Up: Logic Knowledge Representation Previous: Logic Knowledge Next: Isa and instance relationships Up: Predicate logic Previous: Predicate logic
Representation
An example
Predicate logic
Consider the following:

! Prince is a mega star.

! Mega stars are rich.
! Rich people have fast cars.
! An example ! Fast cars consume a lot of petrol.
! Isa and instance relationships
and try to draw the conclusion: Prince's car consumes a lot of petrol.

So we can translate Prince is a mega star into: mega_star(prince) and Mega stars are rich into: m:
mega_star(m) rich(m)
[email protected]
Rich people have fast cars, the third axiom is more difficult:

! Is cars a relation and therefore car(c,m) says that case c is m's car. OR
! Is cars a function? So we may have car_of(m).

Assume cars is a relation then axiom 3 may be written: c,m: car(c,m) rich(m) fast(c).

The fourth axiom is a general statement about fast cars. Let consume(c) mean that car c consumes a
lot of petrol. Then we may write: c: fast(c) m:car(c,m) consume(c) .

Is this enough? NO! -- Does prince have a car? We need the car_of function after all (and addition
to car): c:car(car_of(m),m). The result of applying car_of to m is m's car. The final set of
predicates is: mega_star(prince) m: mega_star(m) rich(m) c:car(car_of(m),m). c,m: car
(c,m) rich(m) fast(c). c: fast(c) m:car(c,m) consume(c) . Given this we could
conclude: consume(car_of(prince)).

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node45.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node46.html 28/08/01

Isa and instance relationships Page 1 of 1 Applications and extensions Page 1 of 1

Next: Applications and extensions Up: Predicate logic Previous: An example Next: Further reading Up: Logic Knowledge Representation Previous: Isa and instance
relationships
Isa and instance relationships
Applications and extensions
Two attributes isa and instance play an important role in many aspects of knowledge representation.
! First order logic basically extends predicate calculus to allow:
The reason for this is that they support property inheritance. " functions -- return objects not just TRUE/FALSE.

" equals predicate added.

isa ! Problem solving and theorem proving -- large application areas.
-- used to show class inclusion, e.g. isa(mega_star,rich). ! STRIPS robot planning system employs a first order logic system to enhance its means-ends
analysis (GPS) planning. This amalgamation provided a very powerful heuristic search.
instance ! Question answering systems.
-- used to show class membership, e.g. instance(prince,mega_star).

From the above it should be simple to see how to represent these in predicate logic.

[email protected]

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node47.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node48.html 28/08/01

Further reading Page 1 of 1 Exercises Page 1 of 1

Next: Exercises Up: Logic Knowledge Representation Previous: Applications and extensions Next: Procedural Knowledge Representations Up: Logic Knowledge Representation Previous:
Further reading
Further reading
Exercises
Artificial Intelligence by Rich and Knight covers this topic well.
1. Assume the following facts:
Knowledge Representation: An AI perspective by H. Reichgelt, Ablex, New York covers all aspects " Steve only likes easy courses.

of this topic with good examples. " Computing courses are hard.
" All courses in Sociology are easy.

The Handbook of Artificial Intelligence (Vol. 1, Ch 3) provides a good concise treatment of this area. " ``Society is evil'' is a sociology course.

See also the Expert Systems course. Represent these facts in predicate logic and answer the question?

What course would Steve like?

2. Find out what knowledge representation schemes are used in the STRIPS system.
[email protected]

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node49.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node50.html 28/08/01

Procedural Knowledge Representations Page 1 of 1 Declarative or Procedural? Page 1 of 1

Next: Declarative or Procedural? Up: Artificial Intelligence II Previous: Exercises Next: An Example Up: Procedural Knowledge Representations Previous: Procedural Knowledge
Representations

Procedural Knowledge Representations Declarative or Procedural?

Declarative knowledge representation:

! Static representation -- knowledge about objects, events etc. and their relationships and states
! Declarative or Procedural? given.
! An Example ! Requires a program to know what to do with knowledge and how to do it.
! Representing How to Use Knowledge
! Further reading Procedural representation:
! Exercises
! control information necessary to use the knowledge is embedded in the knowledge itself. e.g.
how to find relevant facts, make inferences etc.
! Requires an interpreter to follow instructions specified in knowledge.

[email protected]

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node51.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node52.html 28/08/01

An Example Page 1 of 1 Representing How to Use Knowledge Page 1 of 1

Next: Representing How to Use Up: Procedural Knowledge Representations Previous: Declarative Next: Further reading Up: Procedural Knowledge Representations Previous: An Example
or Procedural?
Representing How to Use Knowledge
An Example
Need to represent how to control the processing:
Let us consider what knowledge an alphabetical sorter would need:
direction
! Implicit knowledge that A comes before B etc. -- indicate the direction an implication could be used. E.g. To prove something can fly show it
! This is easy -- really integer comparison of (ASCII) codes for . is a bird. fly(x) bird(x).
" All programs contain procedural knowledge of this sort.
! The procedural information here is that knowledge of how to alphabetise is represented Knowledge to achieve goal
explicitly in the alphabetisation procedure. -- specify what knowledge might be needed to achieve a specific goal. For example to prove
" A declarative system might have to have explicit facts like A comes before B, B comes
something is a bird try using two facts has_wings and has_feathers to show it.
before C etc..

[email protected]
[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node53.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node54.html 28/08/01

Further reading Page 1 of 1 Exercises Page 1 of 1

Next: Exercises Up: Procedural Knowledge Representations Previous: Representing How to Use Next: Weak Slot and Filler Up: Procedural Knowledge Representations Previous: Further reading

Further reading Exercises

Knowledge Representation: An AI perspective by H. Reichgelt, Ablex, New York covers all aspects 1. Discuss how procedural methods may be used to solve the following problems:
of this topic with good examples. " Natural Language Understanding

" The Games of Nim and Kalah (see last years notes for rules).

The Handbook of Artificial Intelligence (Vol. 1, Ch 3) provides a good concise treatment of this area. " Path Planning type tasks.

Winograd's article in Bobrow and Collins deal with fundamental issues and also their robot planner
in Representation and Understanding, Academic Press (1975) deals with declarative verses
procedural issues.
[email protected]

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node55.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node56.html 28/08/01

Weak Slot and Filler Structures Page 1 of 1 Why use this data structure? Page 1 of 1

Next: Why use this data Up: Artificial Intelligence II Previous: Exercises Next: Semantic Nets Up: Weak Slot and Filler Previous: Weak Slot and Filler

Weak Slot and Filler Structures Why use this data structure?
! It enables attribute values to be retrieved quickly
We have already met this type of structure when discussing inheritance in the last lecture. We will " assertions are indexed by the entities
now study this in more detail.
" binary predicates are indexed by first argument. E.g. team(Mike-Hall , Cardiff).

! Properties of relations are easy to describe .

! It allows ease of consideration as it embraces aspects of object oriented programming.

So called because:
! Why use this data structure?
! Semantic Nets ! A slot is an attribute value pair in its simplest form.
" Representation in a Semantic Net
! A filler is a value that a slot can take -- could be a numeric, string (or any data type) value or a
" Inference in a Semantic Net
pointer to another slot.
" Extending Semantic Nets
! A weak slot and filler structure does not consider the content of the representation.
! Frames
" Frame Knowledge Representation
We will study two types:
" Interpreting frames

! Further Reading ! Semantic Nets.

! Exercises ! Frames.

[email protected] [email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node57.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node58.html 28/08/01

Semantic Nets Page 1 of 1 Representation in a Semantic Net Page 1 of 2

Next: Representation in a Semantic Up: Weak Slot and Filler Previous: Why use this data Next: Inference in a Semantic Up: Semantic Nets Previous: Semantic Nets

Semantic Nets Representation in a Semantic Net

The major idea is that:

! The meaning of a concept comes from its relationship to other concepts, and that, The physical attributes of a person can be represented as in Fig. 9.
! The information is stored by interconnecting nodes with labelled arcs.

! Representation in a Semantic Net

! Inference in a Semantic Net
! Extending Semantic Nets

Fig. 9 A Semantic Network

These values can also be represented in logic as: isa(person, mammal), instance(Mike-Hall, person)
[email protected] team(Mike-Hall, Cardiff)

We have already seen how conventional predicates such as lecturer(dave) can be written as instance
(dave, lecturer) Recall that isa and instance represent inheritance and are popular in many
knowledge representation schemes. But we have a problem: How we can have more than 2 place
predicates in semantic nets? E.g. score(Cardiff, Llanelli, 23-6) Solution:

! Create new nodes to represent new objects either contained or alluded to in the knowledge,
game and fixture in the current example.
! Relate information to nodes and fill up slots (Fig: 10).

Fig. 10 A Semantic Network for n-Place Predicate

As a more complex example consider the sentence: John gave Mary the book. Here we have several
aspects of an event.

http://www.cs.cf.ac.uk/Dave/AI2/node59.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node60.html 28/08/01

Representation in a Semantic Net Page 2 of 2 Inference in a Semantic Net Page 1 of 2

Next: Extending Semantic Nets Up: Semantic Nets Previous: Representation in a Semantic

Inference in a Semantic Net

Basic inference mechanism: follow links between nodes.

Two methods to do this:

Intersection search
-- the notion that spreading activation out of two nodes and finding their intersection finds
Fig. 11 A Semantic Network for a Sentence relationships among objects. This is achieved by assigning a special tag to each visited node.

Many advantages including entity-based organisation and fast parallel implementation.

However very structured questions need highly structured networks.

[email protected] Inheritance
-- the isa and instance representation provide a mechanism to implement this.

Inheritance also provides a means of dealing with default reasoning. E.g. we could represent:

! Emus are birds.

! Typically birds fly and have wings.
! Emus run.

in the following Semantic net:

Fig. 12 A Semantic Network for a Default Reasoning

In making certain inferences we will also need to distinguish between the link that defines a new
entity and holds its value and the other kind of link that relates two existing entities. Consider the
example shown where the height of two people is depicted and we also wish to compare them.

We need extra nodes for the concept as well as its value.

Fig. 12 Two heights

http://www.cs.cf.ac.uk/Dave/AI2/node60.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node61.html 28/08/01

Inference in a Semantic Net Page 2 of 2 Extending Semantic Nets Page 1 of 2

Special procedures are needed to process these nodes, but without this distinction the analysis would
be very limited.
Next: Frames Up: Semantic Nets Previous: Inference in a Semantic

Extending Semantic Nets

Here we will consider some extensions to Semantic nets that overcome a few problems (see
Exercises) or extend their expression of knowledge.

Partitioned Networks Partitioned Semantic Networks allow for:

! propositions to be made without commitment to truth.

! expressions to be quantified.

Fig. 12 Comparison of two heights Basic idea: Break network into spaces which consist of groups of nodes and arcs and regard each
space as a node.

Consider the following: Andrew believes that the earth is flat. We can encode the proposition the
earth is flat in a space and within it have nodes and arcs the represent the fact (Fig. 15). We can the
[email protected] have nodes and arcs to link this space the the rest of the network to represent Andrew's belief.

Fig. 12 Partitioned network

Now consider the quantified expression: Every parent loves their child To represent this we:

! Create a general statement, GS, special class.

! Make node g an instance of GS.
! Every element will have at least 2 attributes:
" a form that states which relation is being asserted.

" one or more forall ( ) or exists ( ) connections -- these represent universally quantifiable

variables in such statements e.g. x, y in parent(x) : child(y) loves(x,y)

Here we have to construct two spaces one for each x,y. NOTE: We can express variables as
existentially qualified variables and express the event of love having an agent p and receiver b for
every parent p which could simplify the network (See Exercises).

Also If we change the sentence to Every parent loves child then the node of the object being acted on

http://www.cs.cf.ac.uk/Dave/AI2/node61.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node62.html 28/08/01

Extending Semantic Nets Page 2 of 2 Frames Page 1 of 1

(the child) lies outside the form of the general statement. Thus it is not viewed as an existentially
qualified variable whose value may depend on the agent. (See Exercises and Rich and Knight book
for examples of this) So we could construct a partitioned network as in Fig. 16 Next: Frame Knowledge Representation Up: Weak Slot and Filler Previous: Extending Semantic
Nets

Frames
Frames can also be regarded as an extension to Semantic nets. Indeed it is not clear where the
distinction between a semantic net and a frame ends. Semantic nets initially we used to represent
labelled connections between objects. As tasks became more complex the representation needs to be
more structured. The more structured the system it becomes more beneficial to use frames. A frame
is a collection of attributes or slots and associated values that describe some real world entity.
Frames on their own are not particularly helpful but frame systems are a powerful way of encoding
information to support reasoning. Set theory provides a good basis for understanding frame systems.
Each frame represents:

! a class (set), or
! an instance (an element of a class).

! Frame Knowledge Representation

! Interpreting frames
Fig. 12 Partitioned network

[email protected]
Next: Frames Up: Semantic Nets Previous: Inference in a Semantic

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node62.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node63.html 28/08/01

Frame Knowledge Representation Page 1 of 5 Frame Knowledge Representation Page 2 of 5

Cardinality:
Next: Interpreting frames Up: Frames Previous: Frames
Team-size: 15
Frame Knowledge Representation Coach:

Consider the example first discussed in Semantics Nets (Section 6.2.1):

Person

isa: Mammal Figure: A simple frame system

Cardinality: Here the frames Person, Adult-Male, Rugby-Player and Rugby-Team are all classes and the frames
Robert-Howley and Cardiff-RFC are instances.
Adult-Male

isa: Person Note

Cardinality: ! The isa relation is in fact the subset relation.

! The instance relation is in fact element of.
Rugby-Player ! The isa attribute possesses a transitivity property. This implies: Robert-Howley is a Back and a
isa: Adult-Male Back is a Rugby-Player who in turn is an Adult-Male and also a Person.
! Both isa and instance have inverses which are called subclasses or all instances.
Cardinality: ! There are attributes that are associated with the class or set such as cardinality and on the other
hand there are attributes that are possessed by each member of the class or set.
Height:

Weight: DISTINCTION BETWEN SETS AND INSTANCES

Position: It is important that this distinction is clearly understood.

Team:
Cardiff-RFC can be thought of as a set of players or as an instance of a Rugby-Team.
Team-Colours:
If Cardiff-RFC were a class then
Back
! its instances would be players
isa: Rugby-Player
! it could not be a subclass of Rugby-Team otherwise its elements would be members of Rugby-
Cardinality: Team which we do not want.

Tries: Instead we make it a subclass of Rugby-Player and this allows the players to inherit the correct
properties enabling us to let the Cardiff -RFC to inherit information about teams.
Mike-Hall

instance: Back This means that Cardiff -RFC is an instance of Rugby-Team.

Height: 6-0 BUT There is a problem here:

Position: Centre
! A class is a set and its elements have properties.
Team: Cardiff-RFC ! We wish to use inheritance to bestow values on its members.
! But there are properties that the set or class itself has such as the manager of a team.
Team-Colours: Black/Blue
This is why we need to view Cardiff-RFC as a subset of one class players and an instance of teams.
Rugby-Team
We seem to have a CATCH 22. Solution: MetaClasses
isa: Team

http://www.cs.cf.ac.uk/Dave/AI2/node64.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node64.html 28/08/01

Frame Knowledge Representation Page 3 of 5 Frame Knowledge Representation Page 4 of 5

A metaclass is a special class whose elements are themselves classes. A slot is a relation that maps from its domain of classes to its range of values.

Now consider our rugby teams as: A relation is a set of ordered pairs so one relation is a subset of another.

Since slot is a set the set of all slots can be represent by a metaclass called Slot, say.

Consider the following:

SLOT

isa: Class

instance: Class

domain:

range:

range-constraint:

definition:

default:

to-compute:

single-valued:

Coach

instance: SLOT

domain: Rugby-Team

range: Person

range-constraint: (experience x.manager)

Figure: A Metaclass frame system default:

The basic metaclass is Class, and this allows us to single-valued: TRUE

! define classes which are instances of other classes, and (thus)

! inherit properties from this class.
Colour

Inheritance of default values occurs when one element or class is an instance of a class. instance: SLOT

Slots as Objects domain: Physical-Object

range: Colour-Set
How can we to represent the following properties in frames?
single-valued: FALSE
! Attributes such as weight, age be attached and make sense.
! Constraints on values such as age being less than a hundred Team-Colours
! Default values
instance: SLOT
! Rules for inheritance of values such as children inheriting parent's names
! Rules for computing values isa: Colour
! Many values for a slot.
domain: team-player

http://www.cs.cf.ac.uk/Dave/AI2/node64.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node64.html 28/08/01

Frame Knowledge Representation Page 5 of 5 Interpreting frames Page 1 of 1

range: Colour-Set
Next: Further Reading Up: Frames Previous: Frame Knowledge Representation
range-constraint: not Pink

single-valued: FALSE Interpreting frames

Position A frame system interpreter must be capable of the following in order to exploit the frame slot
representation:
instance: SLOT

domain: Rugby-Player ! Consistency checking -- when a slot value is added to the frame relying on the domain
attribute and that the value is legal using range and range constraints.
range: { Back, Forward, Reserve } ! Propagation of definition values along isa and instance links.
! Inheritance of default. values along isa and instance links.
to-compute: x.position ! Computation of value of slot as needed.
single-valued: TRUE ! Checking that only correct number of values computed.

See Exercises for further instances of drawing inferences etc. from frames.
NOTE the following:

! Instances of SLOT are slots

! Associated with SLOT are attributes that each instance will inherit.
! Each slot has a domain and range. [email protected]
! Range is split into two parts one the class of the elements and the other is a constraint which is
a logical expression if absent it is taken to be true.
! If there is a value for default then it must be passed on unless an instance has its own value.
! The to-compute attribute involves a procedure to compute its value. E.g. in Position where we
use the dot notation to assign values to the slot of a frame.
! Transfers through lists other slots from which values can be derived from inheritance.

Next: Interpreting frames Up: Frames Previous: Frames

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node64.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node65.html 28/08/01

Further Reading Page 1 of 1 Exercises Page 1 of 1

Next: Exercises Up: Weak Slot and Filler Previous: Interpreting frames Next: Strong Slot and Filler Up: Weak Slot and Filler Previous: Further Reading

Further Reading Exercises

Artificial Intelligence by Rich and Knight covers this topic well. 1. Construct Semantic Net representations of the following:

Knowledge Representation: An AI perspective by H. Reichgelt, Ablex, New York covers all aspects 1. Dave is Welsh, Dave is a Lecturer.
of this topic with good examples. 2. Paul leant his new Frank Zappa CD to his best friend.
2. Find out how the SNePS system aims to improve the expressiveness of semantic nets.
The Handbook of Artificial Intelligence (Vol. 1, Ch 3) provides a good concise treatment of this area. 3. Represent the following in a Semantic Net and Comment on how the SNePS system helps.
Volume 3 Chapter 11 also gives good background information on Quinlan's semantic memory
system. The HAM, ACT and MEMOD systems are also dealt with. 1. Mike and Mary's telephone number is the same.
2. John believes that Mike and Mary's telephone number is the same.
Artificial Intelligence Programming by Charniak et al. gives Example LISP code for constructing 4. Represent the following in partitioned semantic networks:
Semantic nets and forms.
1. Every player kicked a ball.
See also the Expert Systems course. 2. All players like the referee.
3. Andrew believes that there is a fish with lungs.
5. Simplify Fig. 16 that represents parent(x) : baby(y) loves(x,y)

so that we can express variables as existentially qualified variables and express the event of
[email protected]
love having an agent p and receiver b for every parent p.

6. Pick a problem area and represent the knowledge in frame based system.
7. Devise algorithms that enable reasoning with frames. Discuss how

1. Inference through inheritance can be achieved.

2. Matching can be achieved.
8. What are the advantages of a frame based knowledge representation?
9. What problems do you envisage a a frame based knowledge representation having? Give
examples of knowledge hard to represent in a frame. How could some problems be overcome?

HINT: address issues to do with semantics, inheritance, expressiveness.

10. What programming languages would be suited to implement a semantic network and frames?

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node66.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node67.html 28/08/01

Strong Slot and Filler Structures Page 1 of 1 Conceptual Dependency (CD) Page 1 of 5

Next: Conceptual Dependency (CD) Up: Artificial Intelligence II Previous: Exercises Next: Scripts Up: Strong Slot and Filler Previous: Strong Slot and Filler

Strong Slot and Filler Structures Conceptual Dependency (CD)

Conceptual Dependency originally developed to represent knowledge acquired from natural
Strong Slot and Filler Structures typically:
language input.
! Represent links between objects according to more rigid rules.
The goals of this theory are:
! Specific notions of what types of object and relations between them are provided.
! Represent knowledge about common situations.
! To help in the drawing of inference from sentences.
! To be independent of the words used in the original input.
! That is to say: For any 2 (or more) sentences that are identical in meaning there should be
only one representation of that meaning.

! Conceptual Dependency (CD) It has been used by many programs that portend to understand English (MARGIE, SAM, PAM). CD
! Scripts developed by Schank et al. as were the previous examples.
! CYC
! Further Reading CD provides:
! Exercises
! a structure into which nodes representing information can be placed
! a specific set of primitives
! at a given level of granularity.

[email protected] Sentences are represented as a series of diagrams depicting actions using both abstract and real
physical situations.

! The agent and the objects are represented

! The actions are built up from a set of primitive acts which can be modified by tense.

Examples of Primitive Acts are:

ATRANS
-- Transfer of an abstract relationship. e.g. give.
PTRANS
-- Transfer of the physical location of an object. e.g. go.
PROPEL
-- Application of a physical force to an object. e.g. push.
MTRANS
-- Transfer of mental information. e.g. tell.
MBUILD
-- Construct new information from old. e.g. decide.
SPEAK
-- Utter a sound. e.g. say.

ATTEND
-- Focus a sense on a stimulus. e.g. listen, watch.
MOVE
-- Movement of a body part by owner. e.g. punch, kick.
GRASP
-- Actor grasping an object. e.g. clutch.
INGEST
-- Actor ingesting an object. e.g. eat.

http://www.cs.cf.ac.uk/Dave/AI2/node68.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node69.html 28/08/01

Conceptual Dependency (CD) Page 2 of 5 Conceptual Dependency (CD) Page 3 of 5

EXPEL -- transition
-- Actor getting rid of an object from body. e.g. ????.
-- start transition
Six primitive conceptual categories provide building blocks which are the set of allowable
dependencies in the concepts in a sentence:
-- finished transition
PP k
-- Real world objects. -- continuing
ACT ?
-- Real world actions. -- interrogative
PA /
-- Attributes of objects. -- negative
AA delta
-- Attributes of actions. -- timeless
T c
-- Times. -- conditional
LOC the absence of any modifier implies the present tense.
-- Locations.
So the past tense of the above example:
How do we connect these things together?
John gave Mary a book becomes:
Consider the example:

John gives Mary a book

The has an object (actor), PP and action, ACT. I.e. PP ACT. The triplearrow ( ) is also a
two link but between an object, PP, and its attribute, PA. I.e. PP PA.

! Arrows indicate the direction of dependency. Letters above indicate certain relationships: It represents isa type dependencies. E.g
o
-- object. Dave lecturerDave is a lecturer.
R
-- recipient-donor. Primitive states are used to describe many state descriptions such as height, health, mental state,
I physical state.
-- instrument e.g. eat with a spoon.
D There are many more physical states than primitive actions. They use a numeric scale.
-- destination e.g. going home.
E.g. John height(+10) John is the tallest John height(< average) John is short Frank Zappa
! Double arrows ( ) indicate two-way links between the actor (PP) and action (ACT). health(-10) Frank Zappa is dead Dave mental_state(-10) Dave is sad Vase
! The actions are built from the set of primitive acts (see above). physical_state(-10) The vase is broken
" These can be modified by tense etc.
You can also specify things like the time of occurrence in the relation ship.
The use of tense and mood in describing events is extremely important and schank
introduced the following modifiers: For Example: John gave Mary the book yesterday

p
-- past
f
-- future
t

http://www.cs.cf.ac.uk/Dave/AI2/node69.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node69.html 28/08/01

Conceptual Dependency (CD) Page 4 of 5 Conceptual Dependency (CD) Page 5 of 5

Now let us consider a more complex sentence: Since smoking can kill you, I stopped Lets look at ! Impossible or difficult to find correct set of primitives.
how we represent the inference that smoking can kill: ! A lot of inference may still be required.
! Representations can be complex even for relatively simple actions. Consider:
! Use the notion of one to apply the knowledge to.
! Use the primitive act of INGESTing smoke from a cigarette to one. Dave bet Frank five pounds that Wales would win the Rugby World Cup.
! Killing is a transition from being alive to dead. We use triple arrows to indicate a transition
from one state to another. Complex representations require a lot of storage
! Have a conditional, c causality link. The triple arrow indicates dependency of one concept on
another. Applications of CD:

MARGIE
(Meaning Analysis, Response Generation and Inference on English) -- model natural language
understanding.
SAM
(Script Applier Mechanism) -- Scripts to understand stories. See next section.
PAM
(Plan Applier Mechanism) -- Scripts to understand stories.

Schank et al. developed all of the above.

Next: Scripts Up: Strong Slot and Filler Previous: Strong Slot and Filler
To add the fact that I stopped smoking
[email protected]
! Use similar rules to imply that I smoke cigarettes.
! The qualification attached to this dependency indicates that the instance INGESTing
smoke has stopped.

Advantages of CD:

! Using these primitives involves fewer inference rules.

! Many inference rules are already represented in CD structure.
! The holes in the initial structure help to focus on the points still to be established.

Disadvantages of CD:

! Knowledge must be decomposed into fairly low level primitives.

http://www.cs.cf.ac.uk/Dave/AI2/node69.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node69.html 28/08/01

Scripts Page 1 of 3 Scripts Page 2 of 3

! Bank Manager, O.
! Policeman, P.
Next: CYC Up: Strong Slot and Filler Previous: Conceptual Dependency (CD)
The Entry Conditions might be:
Scripts
! S is poor.
A script is a structure that prescribes a set of circumstances which could be expected to follow on ! S is destitute.
from one another.
The Results might be:
It is similar to a thought sequence or a chain of situations which could be anticipated.
! S has more money.
It could be considered to consist of a number of slots or frames but with more specialised roles. ! O is angry.
! M is in a state of shock.
Scripts are beneficial because: ! P is shot.

! Events tend to occur in known runs or patterns. There are 3 scenes: obtaining the gun, robbing the bank and the getaway.
! Causal relationships between events exist.
! Entry conditions exist which allow an event to take place The full Script could be described in Fig 19.
! Prerequisites exist upon events taking place. E.g. when a student progresses through a degree
scheme or when a purchaser buys a house.

The components of a script include:

Entry Conditions
-- these must be satisfied before events in the script can occur.
Results
-- Conditions that will be true after events in script occur.
Props
-- Slots representing objects involved in events.
Roles
-- Persons involved in the events.
Track
-- Variations on the script. Different tracks may share components of the same script.
Scenes
-- The sequence of events that occur. Events are represented in conceptual dependency form.

Scripts are useful in describing certain situations such as robbing a bank. This might involve:

! Getting a gun.
! Hold up a bank.
! Escape with the money.

Here the Props might be

! Gun, G.
! Loot, L.
! Bag, B
! Get away car, C.

The Roles might be:

! Robber, S.
! Cashier, M.

http://www.cs.cf.ac.uk/Dave/AI2/node70.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node70.html 28/08/01

Scripts Page 3 of 3 CYC Page 1 of 2

Fig. 12 Simplified Bank Robbing Script

Some additional points to note on Scripts: Next: Further Reading Up: Strong Slot and Filler Previous: Scripts

! If a particular script is to be applied it must be activated and the activating depends on its CYC
significance.
! If a topic is mentioned in passing then a pointer to that script could be held. What is CYC?
! If the topic is important then the script should be opened.
! The danger lies in having too many active scripts much as one might have too many windows ! An ambitious attempt to form a very large knowledge base aimed at capturing commonsense
open on the screen or too many recursive calls in a program. reasoning.
! Provided events follow a known trail we can use scripts to represent the actions involved and ! Initial goals to capture knowledge from a hundred randomly selected articles in the
use them to answer detailed questions. EnCYClopedia Britannica.
! Different trails may be allowed for different outcomes of Scripts ( e.g. The bank robbery goes ! Both Implicit and Explicit knowledge encoded.
wrong). ! Emphasis on study of underlying information (assumed by the authors but not needed to tell to
the readers.
Advantages of Scripts:
Example: Suppose we read that Wellington learned of Napoleon's death
! Ability to predict events.
! A single coherent interpretation may be build up from a collection of observations. Then we (humans) can conclude Napoleon never new that Wellington had died.
Disadvantages: How do we do this?
! Less general than frames. We require special implicit knowledge or commonsense such as:
! May not be suitable to represent all kinds of knowledge.
! We only die once.
! You stay dead.
! You cannot learn of anything when dead.
Next: CYC Up: Strong Slot and Filler Previous: Conceptual Dependency (CD) ! Time cannot go backwards.

[email protected] Why build large knowledge bases:

Brittleness
-- Specialised knowledge bases are brittle. Hard to encode new situations and non-graceful
degradation in performance. Commonsense based knowledge bases should have a firmer
foundation.
Form and Content
-- Knowledge representation may not be suitable for AI. Commonsense strategies could point
out where difficulties in content may affect the form.
Shared Knowledge
-- Should allow greater communication among systems with common bases and assumptions.

How is CYC coded?

! By hand.
! Special CYCL language:
" LISP like.
" Frame based

" Multiple inheritance

" Slots are fully fledged objects.

" Generalised inheritance -- any link not just isa and instance.

http://www.cs.cf.ac.uk/Dave/AI2/node70.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node71.html 28/08/01

CYC Page 2 of 2 Further Reading Page 1 of 1

[email protected]
Next: Exercises Up: Strong Slot and Filler Previous: CYC

Further Reading
Artificial Intelligence by Rich and Knight covers these topic well. In particular see the sections on
scripts for more detail on more complicated scripts with varying tracks. And also a more
comprehensive treatment of CYC.

Knowledge Representation: An AI perspective by H. Reichgelt, Ablex, New York covers all aspects
of this topic with good examples.

The Essentials of AI by Ginsberg provides some useful background information

The Handbook of Artificial Intelligence (Vol. 1, Ch 3) provides a good concise treatment of this area.
Details of MARGIE, SAM and PAM may also be found in Vol. 1, Ch 4.

Examples of how MARGIE and SAM is coded in LISP is given in AI with Common Lisp by Noyes.

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node71.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node72.html 28/08/01

Exercises Page 1 of 1 Reasoning with Uncertainty: Non-Monotonic Reasoning Page 1 of 1

Next: Reasoning with Uncertainty: Non-Monotonic Up: Strong Slot and Filler Previous: Further Next: What is reasoning? Up: Artificial Intelligence II Previous: Exercises
Reading

Exercises Reasoning with Uncertainty: Non-Monotonic

1. Construct CD representation of the following:
Reasoning
1. John begged Mary for a pencil.
2. Jim stirred his coffee with a spoon.
3. Dave took the book off Jim.
4. On my way home, I stopped to fill my car with petrol.
5. I heard strange music in the woods. ! What is reasoning?
6. Drinking beer makes you drunk. ! How can we reason?
7. John killed Mary by strangling her. ! Uncertain Reasoning?
2. Try capturing the differences between the following in CD:
! Non-Monotonic Reasoning
" Default reasoning
1. John slapped Dave, John punched Dave. " Circumscription
2. Sue likes Prince, Sue adores Prince. " Implementations: Truth Maintenance Systems
3. Rewrite the script given in the lecture so that the Bank robbery goes wrong. ! Further Reading
4. Write a script to allow for both outcome of the Bank robbery: Getaway and going wrong and ! Exercises
getting caught.
5. Write a script for enrolling as a student.
6. Find out about how MARGIE, SAM and PAM are implemented. In particular pay attention to
their reasoning and inference mechanisms with the knowledge.
7. Find out how the CYCL language represents knowledge.
[email protected]
8. What are the two levels of representation in the constraints of CYC?
9. Find out the relevance of Meta-Knowledge in CYC and how it controls the interpretations of
knowledge.
10. Find out what levels of concepts CYC has in its ontology.
1. Where should the following concepts be placed in this ontology
# dog

# court case
# South Wales Echo

# Wales

# Pint of Brains Dark

# The Open Golf Championship.

[email protected]

http://www.cs.cf.ac.uk/Dave/AI2/node73.html 28/08/01 http://www.cs.cf.ac.uk/Dave/AI2/node74.html 28/08/01