Transcript
Page 1: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

Coursework Master’s Coursework Master’s Thesis ProposalThesis Proposal

December 1999December 1999

University of South AustraliaUniversity of South Australia

School of Computer and Information ScienceSchool of Computer and Information Science

Student: David BennStudent: David Benn

Supervisor: Dan CorbettSupervisor: Dan Corbett

Implementing Conceptual Graph Processes Implementing Conceptual Graph Processes

Page 2: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

2

PlanPlan

• Conceptual Graphs

• Mineau’s Processes

• Implementing Processes

• Research Issues

• High-level Timeline

• References

Page 3: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

3

Conceptual GraphsConceptual Graphs• Sowa’s [1984] Conceptual Graphs (CGs)

represent knowledge via:– Concepts– Conceptual relations (relationships between

concepts)

Page 4: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

4

Conceptual GraphsConceptual Graphs• Sowa’s theory provides primitives to

model static knowledge.

• Permits modification of knowledge base (KB) via Canonical Formation Rules, e.g.– Copy– Join– Simplify

Page 5: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

5

Conceptual GraphsConceptual Graphsg1: [Feline: ](attr)[Carnivore]

g2: [Sitting]- (agent)[Cat: Elsie] (location)[Mat]

Join g1 and g2 to give g3

g3: [Sitting]- (agent)[Cat: Elsie]-

(attr)[Carnivore] (location)[Mat]

Feline

Cat Tiger Lion

Page 6: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

6

ProcessesProcesses• Dynamic updating of KB falls outside of

scope of theory.

• Mineau [1998] proposed processes as an extension to overcome this lack.

• Need arose from a R&D project to

capture knowledge about corporate

processes, e.g. fabrication of windows [Gerbe´, Keller & Mineau 1998].

Page 7: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

7

ProcessesProcesses

From [Gerbe´, Keller & Mineau 1998]

• Activity dependencies.• Inputs, outputs.• Precondition: before a

window can be assembled, a fabrication order must have been updated with frames and panes which conform to client order.

• Postconditions: when a frame is built, fabrication order is updated; likewise when a window pane is cut.

Page 8: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

8

ProcessesProcesses• [Gerbe´, Keller & Mineau 1998] outlined

a definition of processes in terms of activities and events but said nothing about mechanism.

• [Mineau 1998] introduced mechanism for processes in which:– Dynamic knowledge represented as state

transitions via pre and post conditions.– Applying postconditions consists of asserting

or retracting CGs in KB.

Page 9: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

9

ProcessesProcesses• Actors and demons are specialisations of

processes.– Actors take concepts and compute

individuals of pre-determined concept type. – Demons take concepts and return arbitrary

concept instances.– Processes take CGs as input and return

arbitrary CGs.

• Mineau emphasises simple process primitives which can be extended.

Page 10: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

10

Processes: mechanismProcesses: mechanism• A process consists of a rule set.

• Each rule is a pre/post condition pair.

• Process engine finds matching rule via preconditions and applies postconditions.

• Process takes ≥ 1 CG parameters, yields ≥ 1 CG outputs, asserts/retracts CGs.

process p (in g1...; out g2...) is { ri = prei;posti ,∀i ∈ 1,n[ ] }

Page 11: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

11

Processes: mechanismProcesses: mechanismA trivial algorithm:L0: int n = 1;

L1: while (n <= 10)

L2: n = n + 1;

Rule 0:

Pre0:[Line: L0](to_do)

Post0:[Integer: 1](value)[Variable:*n][Line: L0](done)[Line: L1](to_do)

[Mineau 1998] gives long partial definition for iterative factorial.

Following shows example rule set.

Trigger: [Line: L0](to_do)

Page 12: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

12

Processes: mechanismProcesses: mechanismRule 1a:

Pre1a:

[Line: L1](to_do)[Boolean: True]< ≤ >-

1-[Variable: *n] 2-[Integer: 10]

Post1a:

[Line: L1](done)[Line: L2](to_do)

actor

A trivial algorithm:L0: int n = 1;

L1: while (n <= 10)

L2: n = n + 1;

Page 13: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

13

Processes: mechanismProcesses: mechanismRule 1b:

Pre1b:

[Line: L1](to_do)[Boolean: False]< ≤ >-

1-[Variable: *n] 2-[Integer:

10]

Post1b:

[Line: L1](done)

A trivial algorithm:L0: int n = 1;

L1: while (n <= 10)

L2: n = n + 1;

Page 14: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

14

Processes: mechanismProcesses: mechanismRule 2:

Pre2:

[Line: L2](to_do)

Post2:

[Variable: *n]-

(value)[Integer: *x]- < + >-

[Variable: *n] [Integer: 1]

[Line: L2](done)[Line: L1](to_do)

A trivial algorithm:L0: int n = 1;

L1: while (n <= 10)

L2: n = n + 1;

Page 15: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

15

Implementing ProcessesImplementing Processes• “…there is only a theory which needs to

be refined, implemented, and applied.” [Mineau 1999b]

• Preliminary thoughts:– Need suitable way of expressing rules and processes

(i.e. a process language).

– Use Java for platform independence

– Use Notio framework [Southey & Linders 1999]• KBs, CGs, type lattices, CFRs, CGIF/LF reading/writing

– Use Mineau & Gerbe´’s [1997] contexts for rules.

Page 16: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

16

Digression: ContextsDigression: Contexts

[Human: Fred] (believes) [Statement: #]

[Entity: Earth]–(attr)[Shape: flat](attr)[Location: center_of_universe]

intention

extension

• Sowa [1984] introduced PROPOSITION concept type where referent is set of CGs.

• Mineau & Gerbe´ [1997] formalised contexts.• Intention constrains CGs in extension. (WoA)• Can be used to structure a KB.• Equations to extract graphs by intention, etc.• See Appendix A for overview of contexts.

Page 17: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

17

Research IssuesResearch Issues• How to represent processes

– Syntactically? Algorithmically?

– Can actors, demons, processes have one syntax & representation? Will latter be inefficient? (see p. 11)

• Retraction mechanism– (to_do) vs (done) in algorithm example– Erasure: replacing any CG with “blank”. [Sowa 1999]

• When should process engine stop?– Specify ≥ 1 rules upon which to stop?

– No more matching preconditions?

– Always run to completion or permit stop & restart?

Page 18: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

18

Research Issues (continued)Research Issues (continued)• Interaction with other systems (e.g. CharGer

[Delugach 1999]): CG import, export.• Behaviour of implementation vs original

conception. What has Mineau missed?• Applicability of process mechanism

– [Mineau 1998] only gave factorial as example.

– Needs to be demonstrated on more complex application, e.g. business processes, something more interesting?

Page 19: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

19

Other Research IssuesOther Research Issues• Automatic generation of rule sets. Difficult.• Soundness and completeness

– Do two postconditions conflict?

– [Mineau 1999a]: constraints on processes

• Conceptual Programming Environment– Logic and imperative programming.

– How does current work fit in with this goal (stated in [Mineau 1998])?

• Relationship to other work on dynamic CGs, dataflow languages, etc?

• Higher order processes?

Page 20: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

20

TimelineTimeline• Literature Review

(February, March)• Scope & features

(April)• Design & implement

(May, June, July)• Examples (August,

September)– Factorial– Business processes etc

• Write thesis (October, November)

Lit. Review

Scope

Implement

Examples

Write thesis

Page 21: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

21

ReferencesReferences[Delugach 1999] Delugach, H. CharGer: A Conceptual Graph Editor, [Accessed Online: December 1999], URL: http://www.cs.uah.edu/~delugach/CharGer/

[Gerbe´, Keller & Mineau 1998] Gerbe´, O., Keller, R., Mineau, G. Conceptual Graphs for Representing Business Processes in Corporate Memories. In Proceedings of the 6th International Conference on Conceptual Structures, pp. 401-415, 1998.

[Mineau & Gerbe´ 1997] Mineau, G. & Gerbe´, O. Contexts: A Formal Definition of Worlds of Assertions. In Proceedings of the 5th International Conference on Conceptual Structures, pp. 80-94, 1997.

[Mineau 1998] Mineau, G. From Actors to Processes: The Representation of Dynamic Knowledge Using Conceptual Graphs. In Proceedings of the 6th International Conference on Conceptual Structures, pp. 65-79, 1998.

[Mineau 1999a] Mineau, G. Constraints on Processes: Essential Elements for the Validation and Execution of Processes. In Proceedings of the 7th International Conference on Conceptual Structures, pp. 66-82, 1999.

Page 22: Coursework Master’s Thesis Proposal December 1999 University of South Australia School of Computer and Information Science Student: David Benn Supervisor:

David Benn, December 1999

22

References (continued)References (continued)[Mineau 1999b] Mineau, G. E-mail correspondence from Guy Mineau to David Benn, 1st December, 1999.

[Southey & Linders 1999] Southey, F. & Linders, J.G., Notio — A Java API for developing CG tools. In Proceedings of the 7th International Conference on Conceptual Structures, 1999.

[Sowa 1984] Sowa, J., F. Conceptual Structures: Information Processing in Mind and Machine, Addison-Wesley, 1984.

[Sowa 1999] Sowa, J.F. et al, Conceptual Graph Standard, draft proposed American National Standard (dpANS) NCITS.T2/98-003, 1999. [Accessed Online: November 1999], URL: http://www.bestweb.net/~sowa/cg/cgdpansw.htm

[Sowa 2000] Sowa, J., F. Knowledge Representation: Logical, Philosophical, and Computational Foundations, Brooks/Cole, 2000.


Top Related