Presentation is loading. Please wait.

Presentation is loading. Please wait.

TM CG Notes Latest: 20120509. Outline Big Picture architecture Topic Maps – CG – Notio Notes – Amine Notes – CharGer Notes – Prolog – Analogy –

Similar presentations


Presentation on theme: "TM CG Notes Latest: 20120509. Outline Big Picture architecture Topic Maps – CG – Notio Notes – Amine Notes – CharGer Notes – Prolog – Analogy –"— Presentation transcript:

1 TM CG Notes Latest: 20120509

2 Outline Big Picture architecture Topic Maps – CG – Notio Notes – Amine Notes – CharGer Notes – Prolog – Analogy –

3 Arch 1 TupleSpaceTopic Map Prolog Engine Analogy Engine Harvest Engine SOLR Server Internet Local network

4 Notio CG Graph Graph ConceptRelation A graph is known by its concepts and relations. It is a container of comments. It does not appear to carry any other identity in Notio. Graph does not extend Node, so there is no provision for nested graphs in Notio In CharGer, Graph extends Concept which extends GNode, … An improvement to Notio would be to have Graph extend Node so that it can have an enclosing graph

5 Notio CG Concept Concept Node ConceptType Graph CoreferenceSet Referent Node is an abstract base class. Concept appears to gain subject identity from its ConceptType, its Referent, and its set of CoreferenceSet objects. There is no single ID value; fetching is based on other methods

6 Notio CG Relation Relation Node RelationType Concept IN OUT Relation is known by its RelationType and its collection of concepts Like Concept, there is no ID value; fetching is by other means.

7 Amine Graph Graph Ontology KnowledgeBase Lexicon Identifier CG ConceptCS CG is just a graph of Concepts. CS is a Node which means it can contain CS objects. Not sure where you can get nested graphs…

8 Amine Concept Concept Node TypeDescriptor Variable coreferent Designator Edge INOUT There is no ID in this system. Subject identity appears to come from type, designator, coreferent, and descriptor. Some of those values may be related to Prolog. A Concept lives in its own graph, as well as being a member of other Graphs by way of CGs, which really means it is contained by KnowledgeBase objects

9 Amine Relation Relation Edge RelationType Concept sourcetarget

10 Amine Conceptual Structure CS Node CS parents children Type Concept Type Relation Type CSRuleIndividual

11 CharGer Architecture GraphObject GNode Concept Actor GEdge Relation RelationLabel TypeLabel ArrowCoref CharGer uses Notio for internal nodes and concepts. The concept of a coref appears in several projects. This would be a coreference link. GraphObject has an ID (long) value. A lot of what is here is for the visual editor. Graph

12 Corefence Label The conceptual graph in Figure 1, which represents the sentence John is going to Boston by bus, can be written in the following form in extended CGIF: [Go *x] [Person: John] [City: Boston] [Bus *y] (Agnt ?x John) (Dest ?x Boston) (Inst ?x ?y) In CGIF, concepts are marked by square brackets, and conceptual relations are marked by parentheses. A character string prefixed with an asterisk, such as *x, marks a defining node, which may be referenced by the same string prefixed with a question mark, ?x. These strings, which are called name sequences in Common Logic, represent coreference labels in CGIF and variables in other versions of logic. Following is the equivalent in CLIF: (exists ((x Go) (y Bus)) (and (Person John) (city Boston) (Agnt x John) (Dest x Boston) (Inst x y) )) cg_hbook.pdf A concept node may contain more than one name or coreference label, such as [: John ?z]. In EGs, that node corresponds to a ligature that links two lines of identity; in CLIF, it corresponds to an equality: (= John z).

13 Coreference Link As another example, Figure 7 shows a CG for the sentence If a cat is on a mat, then it is a happy pet. The dotted line that connects the concept [Cat] to the concept [Pet], which is called a coreference link, indicates that they both refer to the same entity. The Attr relation indicates that the cat, also called a pet, has an attribute, which is an instance of happiness cg_hbook.pdf The coreference link in Figure 7 is shown in CGIF by the defining label *x in the concept [Cat: *x] and the bound label ?x in the concept [Pet: ?x].

14 Coreference Note We can accomplish coreference links by assigning each node a symbol (ID value) and substituting that. – What about casting topics into CG structures? Node id=12345 isA Cat Node id= 234324 isA Mat [CAT: 12345] -> (on) -> [Mat:234324] Observation 1: we are using ID (symbols) rather than labels (names) Topic Map CG Observation 2: CG code tends to put whole concepts into structures rather than symbols Note that Amine made these links

15 Referent and Type Each of the four concepts has a type label, which represents the type of entity the concept refers to: Person, Go, Boston, or Bus. Two of the concepts have names, which identify the referent: John or Boston. Each of the three conceptual relations has a type label that represents the type of relation: agent (Agnt), destination (Dest), or instrument (Inst). The CG as a whole indicates that the person John is the agent of some instance of going, the city Boston is the destination, and a bus is the instrument. Therefore: Referent is a locator (typically a name in a cg) Type is a type That summarizes subject identity as we presently know it. Referent could be “Jack” type = Person, but that may not disambiguate. If we set Referent to a symbolic locator of an actual class, that’s a topic chapter05.pdf [ : ]

16 Referents Source: [4]

17 Quantifiers FOR ALL: every cat is on the mat http://www.jfsowa.com/cg/cgexamp.htm Suggests a need for quantifier symbols: FOR ALL (instances) EXISTS (some instance) SOME (?) NONE (?)

18 Terms Concept Context – A Concept that has one or more Graphs as its Referent [1] Cannon – a framework for knowledge organization (e.g. an English dictionary) [2] Coreference Type Graph Lattice – a poset in which every pair of elements has a supremum (or least upper bound) and an infimum (or greatest lower bound) [2] Negation – Negative Context: bits set in a Node to indicate whether the Graph represented by that Node is negative of positive Node Ontology – a subset of a canon dealing with a particular subject domain [2] Partially Ordered Set (Poset) – a set on which exists a binary relation, which is reflexive, anti-symmetric and transitive [2] Referent

19 Formation Rules Referenced paper speaks in terms of formation rules – These are the algorithms of CG manipulation – They are crucial to what a CG can do – Therefore We must study those methods in various implementations to better understand how the graph works http://www.soe.ucsc.edu/share/technical-reports/1995/ucsc-crl-95-22.pdf

20 Lattices The referenced paper shows the concept lattice organized between upper and lower bounding symbols – Notio’s ConceptTypeHierarchy UNIVERSAL_TYPE_LABEL and ABSURD_TYPE_LABEL [2] – What does Amine do? http://www.soe.ucsc.edu/share/technical-reports/1995/ucsc-crl-95-22.pdf

21 Canon The idea of a canon is to capture the overall structure of concepts and the environment in which they are utilized. (no duplicate elements) [2] Three roles: – defines hierarchies of relationships between concept types and between relation types – defines the relationship between each relation type and its associated concept types – defines the relationships between concept types and their instances, if exist, in the real world

22 References [1] http://www.soe.ucsc.edu/share/technical-reports/1995/ucsc-crl-95-22.pdf [2] http://www.users.on.net/~pnguyen/cgi/IEEE-Tkde-0364-1004-4.pdf [3] http://www.fernuni-hagen.de/ps/pubs/CCG.pdf [4] http://www.fb10.uni-bremen.de/khwagner/wissen/pdf/Komposita.pdf


Download ppt "TM CG Notes Latest: 20120509. Outline Big Picture architecture Topic Maps – CG – Notio Notes – Amine Notes – CharGer Notes – Prolog – Analogy –"

Similar presentations


Ads by Google