Knowledge Representation Questions and Answers

Q: What distinguishes event calculus from situation calculus?

Event calculus uses time points and intervals directly, so concurrent events and events with duration are natural, whereas situation calculus threads a sequence of situations through the description. Preconditions and participants are representable in both. The closed world assumption is often used with event calculus but is not required. Choosing between an explicit timeline and a chain of situations determines how easily concurrency and duration can be expressed, which is usually the deciding consideration.

Q: What is a knowledge graph?

A knowledge graph stores entities as nodes and typed relations as edges, at a scale of billions of assertions, and is used to answer queries and to ground search results. It is the descendant of the semantic network, given a formal basis and industrial scale. Growth plots, rule diagrams and proof structures are different objects. That an idea criticised for lacking semantics returned at enormous scale once the semantics were supplied is a fair summary of the field's history here.

Q: What is a triple in the resource description framework?

RDF expresses everything as subject predicate object triples, which is exactly a labelled edge in a graph and is simple enough to be exchanged between systems that agree on nothing else. Definitions, rules and probabilities are layered on top by other standards. The minimal common structure of a labelled directed edge is what allows data from many sources to be merged without prior agreement about schema, which was the design goal.

Q: Why is the triple format chosen for exchanging knowledge?

Any relational fact can be broken into labelled edges, so two data sets can be merged by concatenating their triples without any schema negotiation. That is the whole design argument. Storage is if anything worse than a table, relational databases store tuples of higher arity, and contradictions merge as freely as anything else. Choosing the smallest common structure, so that agreement is required about as little as possible, is the design principle behind most successful interchange formats.

Q: What does the web ontology language add to the triple format?

OWL supplies a description logic semantics, so a reasoner can compute subsumption, detect inconsistency and infer class membership rather than merely retrieving stored edges. Querying is the role of a separate language, and compression and access control are engineering concerns. Layering inference on top of an interchange format, rather than building it in, is what allows the same data to be used by systems that reason and by systems that merely store.

Q: Why does the web ontology language restrict expressiveness deliberately?

Adopting a description logic fragment ensures that subsumption and consistency remain decidable, and several profiles trade expressiveness against complexity for different applications. Readability, storage and contradiction are not the reason. That a standard intended for enormous data volumes should choose its logic for tractability rather than power is the same trade that runs through the whole subject, made explicit in a published specification.

Q: What is instance checking in a description logic?

Instance checking asks whether the asserted facts entail that a particular individual falls under a concept, which is how a reasoner classifies data rather than merely retrieving it. Non-emptiness, validation and counting are different queries. Together with subsumption it makes up the two standard reasoning services, one about concepts and one about individuals, and keeping them distinct clarifies what a reasoner is being asked to do.

Q: What is the terminological part of a description logic knowledge base?

The TBox holds the concept definitions and the subsumption axioms, while the ABox holds the assertions about individuals, and separating them lets the taxonomy be computed once and reused across changing data. Vocabulary lists and inference rules are different things. Dividing a knowledge base into a slowly changing schema and rapidly changing data is the same separation a database makes, and it has the same practical benefits.

Q: Why separate the terminological and assertional parts?

Classifying the concept hierarchy is expensive and need be done only when the definitions change, whereas instance data arrives continuously and is checked against the existing taxonomy. The same reasoner handles both. Expressibility and standards are not the reason. Separating what is stable from what is volatile, so that expensive computation is done on the stable part, is a design pattern that appears wherever a schema and its data are distinguished.

Q: What is a role in a description logic?

Roles are the binary relations, such as has parent or works for, from which complex concepts are built using restrictions and quantifiers. Concepts are the unary predicates. Event participation is a use of roles rather than the definition, and functional roles are a special case. Building complex concept descriptions out of atomic concepts and roles, rather than asserting them, is what allows the hierarchy to be computed rather than declared.

31.

What distinguishes event calculus from situation calculus?

A.It applies only to events with no participants
B.It requires that the closed world assumption be adopted throughout
C.It represents time explicitly rather than as a sequence of situations
D.It cannot represent actions with preconditions at all

Answer: C

Event calculus uses time points and intervals directly, so concurrent events and events with duration are natural, whereas situation calculus threads a sequence of situations through the description. Preconditions and participants are representable in both. The closed world assumption is often used with event calculus but is not required. Choosing between an explicit timeline and a chain of situations determines how easily concurrency and duration can be expressed, which is usually the deciding consideration.

Discuss in Forum

32.

What is a knowledge graph?

A.A plot showing how a knowledge base grows in size over time
B.A diagram of the inference rules a system applies
C.A graph whose nodes are inference steps in a proof
D.A large graph of entities and relations used at web scale

Answer: D

A knowledge graph stores entities as nodes and typed relations as edges, at a scale of billions of assertions, and is used to answer queries and to ground search results. It is the descendant of the semantic network, given a formal basis and industrial scale. Growth plots, rule diagrams and proof structures are different objects. That an idea criticised for lacking semantics returned at enormous scale once the semantics were supplied is a fair summary of the field's history here.

Discuss in Forum

33.

What is a triple in the resource description framework?

A.A subject, a predicate and an object
B.A concept, a definition and an example
C.A rule, a condition and a conclusion
D.An entity, an attribute and a probability

Answer: A

RDF expresses everything as subject predicate object triples, which is exactly a labelled edge in a graph and is simple enough to be exchanged between systems that agree on nothing else. Definitions, rules and probabilities are layered on top by other standards. The minimal common structure of a labelled directed edge is what allows data from many sources to be merged without prior agreement about schema, which was the design goal.

Discuss in Forum

34.

Why is the triple format chosen for exchanging knowledge?

A.It prevents contradictory statements from being merged
B.It is the simplest structure that can express any relation
C.Triples require less storage than any other representation
D.It matches the internal format of most relational databases

Answer: B

Any relational fact can be broken into labelled edges, so two data sets can be merged by concatenating their triples without any schema negotiation. That is the whole design argument. Storage is if anything worse than a table, relational databases store tuples of higher arity, and contradictions merge as freely as anything else. Choosing the smallest common structure, so that agreement is required about as little as possible, is the design principle behind most successful interchange formats.

Discuss in Forum

35.

What does the web ontology language add to the triple format?

A.A query language for retrieving triples from a store
B.Access control determining who may read each triple
C.Formal semantics permitting inference over the graph
D.A compression scheme reducing the size of large graphs

Answer: C

OWL supplies a description logic semantics, so a reasoner can compute subsumption, detect inconsistency and infer class membership rather than merely retrieving stored edges. Querying is the role of a separate language, and compression and access control are engineering concerns. Layering inference on top of an interchange format, rather than building it in, is what allows the same data to be used by systems that reason and by systems that merely store.

Discuss in Forum

36.

Why does the web ontology language restrict expressiveness deliberately?

A.To make the syntax easier for people to write
B.To reduce the storage required for large ontologies
C.To prevent contradictory ontologies from being published
D.To keep reasoning decidable at web scale

Answer: D

Adopting a description logic fragment ensures that subsumption and consistency remain decidable, and several profiles trade expressiveness against complexity for different applications. Readability, storage and contradiction are not the reason. That a standard intended for enormous data volumes should choose its logic for tractability rather than power is the same trade that runs through the whole subject, made explicit in a published specification.

Discuss in Forum

37.

What is instance checking in a description logic?

A.Deciding whether an individual belongs to a concept
B.Checking that every concept has at least one instance
C.Verifying that the instances stored are syntactically valid
D.Counting how many individuals a concept contains

Answer: A

Instance checking asks whether the asserted facts entail that a particular individual falls under a concept, which is how a reasoner classifies data rather than merely retrieving it. Non-emptiness, validation and counting are different queries. Together with subsumption it makes up the two standard reasoning services, one about concepts and one about individuals, and keeping them distinct clarifies what a reasoner is being asked to do.

Discuss in Forum

38.

What is the terminological part of a description logic knowledge base?

A.The rules by which inference is carried out
B.The definitions of the concepts and roles
C.The assertions about particular individuals
D.The vocabulary of terms without their meanings

Answer: B

The TBox holds the concept definitions and the subsumption axioms, while the ABox holds the assertions about individuals, and separating them lets the taxonomy be computed once and reused across changing data. Vocabulary lists and inference rules are different things. Dividing a knowledge base into a slowly changing schema and rapidly changing data is the same separation a database makes, and it has the same practical benefits.

Discuss in Forum

39.

Why separate the terminological and assertional parts?

A.Assertions cannot be expressed in the same language as definitions
B.The separation is required by the web ontology standard
C.The definitions change slowly while the data changes constantly
D.The two require entirely different inference algorithms

Answer: C

Classifying the concept hierarchy is expensive and need be done only when the definitions change, whereas instance data arrives continuously and is checked against the existing taxonomy. The same reasoner handles both. Expressibility and standards are not the reason. Separating what is stable from what is volatile, so that expensive computation is done on the stable part, is a design pattern that appears wherever a schema and its data are distinguished.

Discuss in Forum

40.

What is a role in a description logic?

A.The part an individual plays in an event
B.A category to which an individual belongs
C.A function assigning a value to an individual
D.A binary relation between individuals

Answer: D

Roles are the binary relations, such as has parent or works for, from which complex concepts are built using restrictions and quantifiers. Concepts are the unary predicates. Event participation is a use of roles rather than the definition, and functional roles are a special case. Building complex concept descriptions out of atomic concepts and roles, rather than asserting them, is what allows the hierarchy to be computed rather than declared.

Discuss in Forum

Loading…

Topics

Other Subjects

Tags: Knowledge Representation Questions and Answers || Knowledge Representation MCQ Questions and Answers || Knowledge Representation GK Questions and Answers || Knowledge Representation GK MCQ Questions || Knowledge Representation Multiple Choice Questions and Answers || Knowledge Representation GK || GK on Knowledge Representation || Artificial Intelligence Questions and Answers || Artificial Intelligence MCQ Questions and Answers || Artificial Intelligence GK Questions and Answers || GK on Artificial Intelligence

Knowledge Representation Questions and Answers

Topics

Other Subjects

About Us

Stay informed about all important Dates & Notifications