group: architecture, frameworks and tools to regulated open mas presented by carolina felicissimo

An Architecture for Open Multi-Agent Systems An Architecture for Open Multi-Agent Systems Ontology-Driven ApplicationsOntology-Driven Applications

An Architecture for Regulated Open Multi-Agent Systems

State of the Art and Open Problems

Group: Architecture, frameworks and tools to regulated open MAS

Presented by Carolina Felicissimo

Carolina Howard Felicíssimo © LES/PUC-Rio

Outline

• Motivation

• State of the Art

• Most Relevant Authors

• Open Problems

• An Architecture for Regulated Open MAS

• Conclusion

• Future Trends


Motivation

• For an Open MAS, how to express the

– Environment’s laws

– Organizations’ laws

– Roles’ laws

– Interactions’ laws

in a consistent and correct way, respecting inheritance?

• Many formalisms’ techniques from Artificial Intelligence

• but the Architecture for regulated open MAS will follow the principles of Software Engineering


Motivation

Examples:

• An Environment’s Law:– Everybody have to obey traffic laws

• An Organization’s Law:– Everybody have to stop the car in red traffic lights

• A Role’s Law:– Everybody is permitted to cross the red traffic lights after 10:00 PM, if

the interactions’ laws are obeyed

• An Interaction’s Law:– No car driver can collide with other car drives

– No car driver can collide with pedestrians


Motivation

• Which is the best formalism for representing knowledge?

It has to:

– Provide a precise characterization of the knowledge base to be used

– Find implicit consequences of its explicit represented knowledge

– Exploit the notion of hierarchical structure

– Be easy to represent

– Provide reasoning efficiency

The more expressive the language, the harder the reasoning !!!

How to balance expressiveness and reasoning?


State of the Art

• Deontic Logic

– Founded by Georg Henrik von Wright in 1951

• Description Logic

– Semantic networks in 1967

– Frame systems in 1981

– Network based-structures in 1992

• Defeasible Logic

– Founded by Donald Nute in 1987

• Defeasible Description Logic

– Guido Governatori. Rules and Rule Markup Languages for the Semantic Web. LNCS 3323, pp. 98-112, 2004


State of the Art

• Deontic Logic

Develop a technique for deciding, whether the propositions it studies are logically true or not -> Decision Problems can be solved

- If an act is permitted, then it is not prohibited• Ex.: Walk in sidewalks is permitted

• If an act is obligatory, then it is permitted and it is not prohibited• Ex.: Drive in roads is obligatory

• If an act is prohibited, then it is not obligatory and it is not permitted• Ex.: Walk in roads is prohibited

• If is not permitted not perform an act, then this act is obligatory• Ex.: Not drive in roads is not permitted


State of the Art

• Deontic Logic

- If is prohibited not perform an act, then this act is obligatory• Ex.: Not drive in roads is prohibited

• Ex.: Not stop in red traffic lights is prohibited

- If is permitted or is obligatory not perform an act, then this act is prohibited• Ex.: Drive in roads is permitted

• Ex.: Not drive in red traffic lights is obligatory

- If is permitted perform and not perform an act, then this act is permitted• Ex.: In a smoking compartment is permit to smoke or not. So, is permitted to

smoke


State of the Art

• Description Logic

– Monotonic formalisms based upon first-order logic

– First order logic is inappropriate for reasoning on partial, incomplete and inconsistent knowledge base, and when there have conflicts

-> can give wrong answers

– Allows for the representation of

• concept conjunctions (Father Man Π Parent)

• concept disjunctions (Person Man Woman)

• concept negations (Woman ¬ Man Π Person)

• value or role restriction constructs ( hasChild.female)


State of the Art

• Defeasible LogicDefeasible: Adj. Capable of being annulled or invalidated

– It is a simple but efficient formalism for nonmonotonic reasoning based on rules and priorities

– Priorities on rules may be used to resolve some conflicts among rules

– Rules may support conflicting conclusions

– Skeptical: Conflicting rules do not fire• Consistency is preserved

– Classical negation is used in the heads and bodies of rules• Negation-as-failure is not used but can be emulated

– It doesn’t have disjunction

– Low computational complexity


State of the Art

• Nonmonotonic Reasoning

Grigoris Antoniou

“Provides formal methods that enable intelligent systems to operate

adequately when faced with incomplete or changing

information.”

Can derive some meaningful solutions in the presence of conflicts in

the knowledge base used

Can deal with rules with exceptions


State of the Art

• The Defeasible theory

– Facts: are indisputable statements. • Ex.: emu(tweety) “Tweety is an emu”

– Strict rules: are rules, whenever the premises are indisputable (e.g., facts) then so is the conclusion.

• Ex.: emu(X) -> bird(X) “Emus are birds”

– Defeasible rules: are rules that can be defeated by contrary evidence• Ex.: bird(X) => flies(X) “Birds typically fly”

– Defeaters: are rules that cannot be used to draw any conclusions. Their only use is to prevent some conclusions. They are used to defeat some defeasible rules by producing evidence to the contrary.

• Ex.: heavy(X) ~> ¬flies(X) “If an animal is heavy then it might not be able to fly”

– A superiority relation among rules: it is used to define priorities among rules, i.e., where one rule may override the conclusion of another rule


State of the Art

• Defeasible Logic

– It has been applied to

• Legal Knowledge

• Legal Reasoning

• Automated Negotiation

• Contracts

• Business Rules

• Multi-agent Systems

• Normative Reasoning

• Defeasible Logic reasoning

– It is a rule-based approach for efficient reasoning with incomplete and inconsistent information

– Is not as strong or conclusive as the reasoning in Description Logic


Defeasible Logic – An Example

• User Requirements & Preferences:

– Carlos is looking for an apartment of at least 45m2 with at least 2 bedrooms. If it is on the 3rd floor or higher, the house must have an elevator. Also, pet animals must be allowed.

– Carlos is willing to pay $300 for a centrally located 45m2 apartment, and $250 for a similar flat in the suburbs. In addition, he is willing to pay an extra $5 per m2 for a larger apartment, and $2 per m2

for a garden.

– He is unable to pay more than $400 in total. If given the choice, he would go for the cheapest option. His 2nd priority is the presence of a garden; lowest priority is additional space.



• Predicates Used in Formalization:

– size(x,y), where y is the size of apartment x (in m2)

– bedrooms(x,y), where apartment x has y bedrooms

– price(x,y), where y is the price for x

– floor(x,y), where apartment x is on the y-th floor

– gardenSize(x,y), where apartment x has a garden of size y

– lift(x), meaning that there is an elevator in the house of x

– pets(x), meaning that pets are allowed in x

– central(x), meaning that x is centrally located

– acceptable(x), meaning that flat x satisfies Carlos’s requirements

– offer(x,y), meaning that Carlos is willing to pay $ y for flat x



• Formalization of Requirements:

– r1: => acceptable(X)

– r2: bedrooms(X,Y), Y < 2 => ¬acceptable(X)

– r3: size(X,Y), Y < 45 => ¬acceptable(X)

– r4: ¬pets(X) => ¬acceptable(X)

– r5: floor(X,Y), Y > 2, ¬lift(X) => ¬acceptable(X)

– r6: price(X,Y), Y > 400 => ¬acceptable(X)

– r2 > r1, r3 > r1, r4 > r1, r5 > r1, r6 > r1

– r7: size(X,Y), Y ≥ 45, garden(X,Z), central(X) => offer(X, 300 + 2Z + 5(Y−45))



• Formalization of Requirements:

– r8: size(X,Y), Y ≥ 45, garden(X,Z),¬central(X) => offer(X, 250 + 2Z + 5(Y−45))

– r9: offer(X,Y), price(X,Z), Y < Z => ¬acceptable(X)

– r9 > r1

– r10: acceptable(X), price(X,Z), not(acceptable(Y), Y ≠ X, price(Y,W), W < Z) => cheapest(X)

– r11: cheapest(X), gardenSize(X,Z), not(cheapest(Y), Y ≠ X, gardenSize(Y,W), W < Z) => largestGarden(X)

– r12: largestGarden(X), size(X,Z), not(largestGarden(Y), Y ≠ X, size(Y,W), W < Z) => rent(X)



App Bed Size Cent Floor Lift Pets Gard Price

a1 1 50 yes 1 no yes 0 300

a2 2 45 yes 0 no yes 0 335

a3 2 65 no 2 no yes 0 350

a4 2 55 no 1 yes no 15 330

a5 3 55 yes 0 no yes 15 350

a6 2 60 yes 3 no no 0 370

a7 3 65 yes 1 no yes 12 375



• Results of User Preferences:

– Apartment a1 is not acceptable because it has one bedroom only (rule r2).

– Apartments a4 and a6 are unacceptable because pets are not allowed (rule r4).

– Apartment a2 is unacceptable because it costs more than the $300 Carlos is willing to pay (rules r7 & r9).

– The rest, a3, a5 and a7, are acceptable.

– Apartments a3 and a5 are the cheapest acceptable apartments (rule r10)

– a5 is selected because it has larger garden than a3 (rules r11 and r12)


State of the Art

• Defeasible Description Logic

– Defeasible assertions are added to description logic

– Reasoning on partial or incomplete knowledge bases, and when there exist conflicts in a decidable way

– Should be used when an open world is assumed


Most Relevant Authors

• Deontic Logic– Georg Henrik von Wright (Founder. 1951)

• Description Logic– Franz Baader, Diego Calvanes, Deborah McGuinnes, Daniele Nardi, and Peter

Patel-Schneider

• Defeasible Logic– Donald Nute (Founder. 1987)

• Defeasible Description Logic – Guido Governatori (from University of Queensland, Australia)

– Grigoris Antoniou (from University of Crete, Greece )

• Normative Multiagent Systems– Guido Boella (from Università di Torino, Italy)

– Leendert van der Torre (from CWI Amsterdam and TU Delft)

• Ontologies– Deborah McGuinnes, Ian Horrocks, Richard Benjamins, Peter Patel-Schneider


Open Problems

• How to balance expressiveness and reasoning?

• Which is the best formalization to be used?

– It depends on the problem that is being studied


The Architecture

• First we have to decide what kind of formalism we will use

• Then, we can think about an architecture for regulated open MAS


Conclusion

• Thousand of different logics !!!

• Each one for specific purposes :/

• You have to find the best one that formalize your problem


Future Trends

• Choose one formalism

• Balance the trade-offs about the choice made

• Structure the paper for the AAMAS-2006 Conference


Bibliography

• Deontic Logic in Computer Science: Normative System Specification (Hardcover). Meyer and Wieringa. 400 pages. February, 1994.

– Available at Amazon for 176 dollars :/

• Defeasible logic. Donald Nute. In Handbook of Logic in Artificial Intelligence and Logic Programming, volume 3, pages 353–395. Oxford University Press, 1987. 528 pages.

– Available at Oxford University Press for 250 dollars ://

• Nonmonotonic Reasoning. Grigoris Antoniou. 275 pages. April, 1997.

– Available at MIT Press for 55 dollars

• The Description Logics Handbook. F. Baader, D. Calvanes, D. McGuinnes, D. Nardi and P. Patel-Schneider, editors. Cambridge University Press, Cambridge, 2003. 574 pages. January, 2003.

– Available at The Cambridge University Press for 85 euros :/

– Or electronic with me without the Bibliography part


Bibliography

• Law and the Semantic Web. Legal Ontologies, Methodologies, Legal Information Retrieval, and Applications. Lecture Notes in Computer Science from Springer. Volume 3369/2005

– Available electronic at http://www.springerlink.com/openurl.asp?genre=issue&issn=0302-9743&volume=3369&issue=preprint

• Deontic logic, agency and normative systems. [Delta] EON'96, Third International Workshop on Deontic Logic in Computer Science / Mark A. Brown and José Carmo (eds.). International Workshop on Deontic Logic in Computer Science (3. 1996.Sesimbra, Portugal)

– Available at PUC-Rio’s library :)


Links

• The University of Queensland, Australia (http://www.itee.uq.edu.au/~guido/Papers/deontic.html)

• Leendert van der Torre Publications (http://homepages.cwi.nl/~torre/papers-c.html)

• Guido Boella Publications

(http://www.di.unito.it/cgi-bin/biblio/bibsearchNL?author=Boella)

group: architecture, frameworks and tools to regulated open mas presented by carolina felicissimo

Documents

carolina felicissimo

traffic laws

art deontic logic

defeasible logic

red traffic lights

art description logic

car driver

relevant authors open