340 likes | 554 Views
ITCS452: Knowledge Representation. Dr. Jarernsri L. Mitrpanont. Knowledge Representations. A number of knowledge-representation techniques have been devised: Semantic nets Frames Scripts Conceptual graphs Logic Rules. Associative (Semantic) Networks.
E N D
ITCS452:Knowledge Representation Dr. Jarernsri L. Mitrpanont
Knowledge Representations A number of knowledge-representation techniques have been devised: • Semantic nets • Frames • Scripts • Conceptual graphs • Logic • Rules
Associative (Semantic) Networks Developed by Quillian 1968 as a graphical represention of knowledge within English sentences. It is represented by a labeled, directed graph. • nodes/objects / concepts • arcs / links relationships / associations The powerof an associative network is derived from the richness of its underlying relationships that provide the cohesive structure for interpreting the structure and inferred new facts.
A Semantic net is a representation in which • Lexically there are nodes, links, and application specific link label. • Structurally, each link connects a tail node to a head node. • Semantically, the nodes and links denote application-specific entities / relationship • show an entity’s attributes and its relationships to other entities. • exhibit any taxonomic structure inherent in a group of entities or concepts. • depict the way in which knowledge is distributed or clustered about entities in KB.
Brachman et al. 1983 • Taxonomic Links in Semantic Networks • Link in semantic Net is considered as the backbone of the hierarchy which provides the “inheritance” link between the representational objects or nodes. • This Link is often called “IS-A” ( also known as “IS,” “SUPERC,” “AKO,” “SUBSET,” etc. )
Link : • The IS-A connections formed a hierarchy of the types being connected. • The hierarchical organization make it easy to distribute “properties” being shared among nodes in the hierarchy. • This organization make the semantic net an efficient storage scheme. • Inheritancepropagates from top node to all nodes below in the hierarchy.
Semantic Nets • A classic representation technique for propositional information • Propositions – a form of declarative knowledge, stating facts (true/false) • Propositions are called “atoms” – cannot be further subdivided. • Semantic nets consist of nodes (objects, concepts, situations) and arcs (relationships between them).
Common Types of Links • IS-A – relates an instance or individual to a generic class • A-KIND-OF – relates generic nodes to generic nodes
Two major types of IS-A relations Generic / generic relations usages • Subset / Superset (fightingships / battleship) • Generalization / Specialization (Restaurant / Fast-food) • predicates + quantifier if x is a Fast-food then x is a Restaurant.
AKO • (an elephant is a kind of mammal.) • Conceptual containment • (A triangle is a polygon.) • Role value restriction • (an elephant trunk is a cylinder 1.3 meters in length.) • Set and its characteristic type. • (an elephant / a set of elephants)
Two major types of IS-A relations Generic / individual relationships • Set membership (Clyde is a camel) • Predication Apply to individual BROWN (camel) • Conceptual containment (King / King of France) • Abstraction (The eagle is an endangered species)
Different dimensions which IS-A vary • By the effects of node meaning • the types of the related items • between two sets (subset) • between predicates • if something is a whale, then it is a mammal • between concepts / descriptions
By the syntactic function of the link • sentence-forming vs. concept-forming • Indian-Elephant is a Brown-thing • Sentence-forming with “quantifier” notion (“every”) • Bird is a Flying-thing. (Universal vs. default) • Person is a Mammal. • Person is a Human-being. • Presence an assertion or not (Believes / Facts) • Apple is-a red fruit.
Air WING IS-A Tweety CANARY Animal bird PENGUIN FLY WALK A Semantic Network A “bird” user How BREATHE FLY HAS IS-A IS-A TRAVEL IS-A How user WALK TRAVEL Operation How do you travel ?
WING BIRD FLY WING CANARY BIRD FLY A Semantic Network A “bird” HAS TRAVEL Expanding the network HAS IS-A TRAVEL
AIR WING Animal CANARY BIRD FLY AIR WING Animal CANARY BIRD Tweety FLY A Semantic Network More general HAS Breathe IS-A IS-A TRAVEL More specific HAS Breathe IS-A IS-A IS-A TRAVEL
AIR WING Animal CANARY BIRD Tweety FLY WALK A Semantic Network Exceptional Handling HAS Breathe IS-A IS-A IS-A IS-A TRAVEL PENGUIN TRAVEL Operation : How do you travel ?
WING AIR Animal CANARY BIRD Tweety FLY WALK A Semantic Network How do you travel ? HAS Breathe IS-A IS-A IS-A user FLY TRAVEL IS-A PENGUIN How do you travel ? user TRAVEL WALK Inheritance
PROLOG and Semantic Nets • In PROLOG, predicate expressions consist of the predicate name, followed by zero or more arguments enclosed in parentheses, separated by commas. • Example: mother(becky,heather) means that becky is the mother of heather
PROLOG Continued • Programs consist of facts and rules in the general form of goals. • General form: p:- p1, p2, …, pN p is called the rule’s head and the pi represents the subgoals • Example: spouse(x,y) :- wife(x,y) x is the spouse of y if x is the wife of y
Object-Attribute-Value Triple • One problem with semantic nets is lack of standard definitions for link names (IS-A, AKO, etc.). • The OAV triplet can be used to characterize all the knowledge in a semantic net.
Problems with Semantic Nets • To represent definitive knowledge, the link and node names must be rigorously defined. • A solution to this is extensible markup language (XML) and ontologies. • Problems also include combinatorial explosion of searching nodes, inability to define knowledge the way logic can, and heuristic inadequacy.
Schemata • Knowledge Structure – an ordered collection of knowledge – not just data. • Semantic Nets – are shallow knowledge structures – all knowledge is contained in nodes and links. • Schema is a more complex knowledge structure than a semantic net. • In a schema, a node is like a record which may contain data, records, and/or pointers to nodes.
Frames • One type of schema is a frame (or script – time-ordered sequence of frames). • Frames are useful for simulating commonsense knowledge. • Semantic nets provide 2-dimensional knowledge; frames provide 3-dimensional. • Frames represent related knowledge about narrow subjects having much default knowledge.
Frames Continued • A frame is a group of slots and fillers that defines a stereotypical object that is used to represent generic / specific knowledge. • Commonsense knowledge is knowledge that is generally known. • Prototypes are objects possessing all typical characteristics of whatever is being modeled. • Problems with frames include allowing unrestrained alteration / cancellation of slots.
condition : x/16 < x < 65 condition : x/16 < x <120 FRAME Definition : A frame is a data structure which represent an entity type. It consists of a collection of a named ‘slots’ each of which can be ‘filled’ by values or pointers to other frames. Employee Frame Family-man Frame named generic : MARRIED disposition named skill marital status disposition number of legs married to aged noschildren condition : x/family woman aged default:two _ _ _ _
Type of slot values Instantiation is a process that the slots of a frame are filled. • Slots values types : • Generic values. • Default values. • Condition • Other frame • Link / pointer
rames Structure F • Frames contain Slots and Slot Values • Frames may describe instances or classes • instance frames describe individual thing • class frames describe entire class of things • special slots • Is - A slots (a member of the class -- ties instances to their classes) • AKO slot (ties classes together -- class - subclasses)
Happy Grumpy Frame and Semantic Net Manager is-a Everything likes ako Dwarfs Manager ako is-a likes Competitors Grumpy Happy ako Managers Managers is-a Blimpy Grumpy Happy Is-a A Simple class hierarchy Happy Likes
class constructor • instance constructor • slot writer • slot reader Frames have access procedures Construct a class frame : Write Happy into the Likes slot : Grumpy Manager Manager Is-a Ako Competitors Happy Likes Happy Construct an instance frame : Read the value in the Likes slot : Grumpy Grumpy Manager Is-a Manager Is-a Likes Happy Likes Happy • Inheritance enables When-Constructed procedures to move Default slot values from classes to instances. • Demons are procedures that are activated if there are changes or uses of the values. • Frame can capture general knowledge about events.
Inference types in Frame Systems • Inferred existence • Inferred generic properties • Default properties • Recognition of abnormal situations • Inference by Analogy
Advantages : • The clustering of properties relating to an entity or entity set is natural in the sense that people are familiar (expert approach). • Frame structure accommodate a taxonomy of knowledge with the highly modular structure. • Generalization and specialization are accommodated via the class/subclass and member of links. • Frames facilitate reasoning. Disadvantages : • There are no formal theory on inference mechanisms of frames.