Argumentation and Trust: Issues and New Challenges Jamal Bentahar Concordia University (Montreal,...

Preview:

Citation preview

Argumentation and Trust: Argumentation and Trust: Issues and New ChallengesIssues and New Challenges

Jamal BentaharConcordia University (Montreal, Canada)

University of Namur, Belgium, University of Namur, Belgium, June 26, 2007June 26, 2007

2

Overview

• Problem and Motivations

• Negotiation Framework

• Trustworthiness Model

• Implementation

• Application Areas

3

Context and Problem

• Multi-agent Systems: interacting autonomous agents

• Communication Protocols: specifying allowed communicative acts

• Open and dynamic MAS need flexible protocols

• Examples: negotiation, deliberation, and persuasion

• Security engineering: a new challenge in agent-based software

engineering

• Distributed setting: e.g. semantic-grid computing

• Computational efficiency

4

Proposed Approaches for Interacting Agents

Mental Approach

Mental Approach

Private states: Beliefs, Desires, Intentions, etc.

Social Approach

Social Approach

Public states: Social

commitments

Argumentative Approach

Argumentative Approach

Argumentation theory +

reasoning

Allen and Perrault, 1980

Cohen and Levesque, 1990

and others

Singh, 2000

Colombetti, 2000

and others

Amgoud and Maudet, 1999

McBurney et al., 2002

and others

5

Motivations

• How to trust negotiating agents within a multi-agent

system:

• Resources sharing and mutual access

Centralized Approaches

Vulnerable to attacks

Vulnerable to attacks Reasoning

CapabilitiesReasoning Capabilities

QuantitativeProbabilistic-based

QuantitativeProbabilistic-based

Decentralized Approach

6

Overview

Problem and Motivations

• Negotiation Framework

• Trustworthiness Model

• Implementation

• Application Areas

7

Agent Architecture

8

Negotiation Framework

Agent 1Agent 1 Agent 2Agent 2

Social Commitments

+

Argumentation

Social Commitments

+

Argumentation

Speech Act Theory + Action Logic

Speech Act Theory + Action Logic

Negotiation

SpecificationSpecification

Reasoning + SemanticsReasoning + Semantics

9

Negotiation Framework

Argumentation Theory

Agent Negotiation

SupportSupport

FlexibilityFlexibility EfficiencyEfficiency

Dialogue Games

Dialogue Games

RelevanceTheory

RelevanceTheory

Logic-based

Reasoning

Logic-based

Reasoning

10

Dialogue Games

• Abstract structures that can be composed:• Sequencing:

• Embedding:

• Parallelization:

• Argumentation-driven decision making process

Game 1 Game 2,

Game1

Game 2… …

Game 1 Game 2//

11

Dialogue Games: Specification

• Initiative / reactive dialogue games

• A simple language

• Cond: generating arguments from the agent’s argumentation system

Action_Ag1 Action_Ag2

Cond

12

Dialogue Games: Specification

ConditionAction_Ag1 Action_Ag2

Create(Ag1,SC(p))

Accept(Ag2,SC(p))

Challenge(Ag2,SC(p))

Refuse(Ag2,SC(p))

Termination

Justification

Persuasion

c1

c2

c3

13

Agent Communication

• Action_Agi {Make-Offer, Make-Counter-

Offer, Withdraw, Satisfy, Violate, Accept,

Refuse, challenge, Justify, Defend, Attack}

Argumentation system

Argumentation system

Communicative Actions

Communicative ActionsSupportsSupports

14

• A powerful framework for interacting agents:

• Making decisions

• Assessing the validity of information

• Resolving differences of opinion

• Argumentation focuses on interactions where parties

plead for and against some conclusion

• Essential ingredient of negotiation, persuasion and

collaborative decision-making.

Argumentation

15

• The notion of argument:

a pair <Premises, Conclusion>

• An argument is a pair (P, c) where P is a

set of beliefs and c is a formula, such

that:

i) P is consistent, ii) P c et iii) P is minimal

Formal Argumentation

16

• Attack relation: binary relation between

arguments

• An argument (P1, c1) attacks another

argument (P2, c2) iff

• c1 c2 or x P2 | c1 x

Argumentation Dynamics

17

Argumentation Dynamics

18

Overview

Problem and Motivations

Negotiation Framework

• Trustworthiness Model

• Implementation

• Application Areas

19

Trust in MAS

• Two approach types into trusting multi-agent systems: centralized

and decentralized

• Centralized approaches: e.g. eBay and Amazon Auctions

• The ratings are stored centrally and summed up to give an overall rating

• Reputation is a global single value

• The model can be unreliable, particularly when some buyers do not

return ratings

• These models are not suitable for applications in open MAS such as

agent negotiation

20

Trust in MAS

• Three main decentralized approaches:

• Building on agents’ direct experiences of

interaction partners

• Using information provided by other agents

• Certified information provided by referees

• Examples: Regret, Referral, Fire

21

• Qualitative approach: Using argumentation to reason about trust

• Quantitative approach: Probability function• Rep : AAD [0, 1]

• Local beliefs

• Global beliefs: testimonies of witnesses

_ ( ) _ ( ) ( )

_ _ ( ) _ _ ( )a a

aa a

Ag Agb bAgb

Ag Agb b

Nb Arg Nb CAg AgRep Ag

T Nb Arg T Nb CAg Ag

Foundation

22

Illustration

23

• Central Limit Theorem and the Law of Large Numbers

• If M > w Then Return True

Else Return False

1

1

( ) ( ) ( ) ( )

( ) ( ) ( )a b b i

b ba

nAg Ag Ag Agi i i bi

nAg Agi i ii Ag

Rep N TR RepAg Ag Ag AgM

Rep N TRAg Ag Ag

Assessing Agent’s Reputation

24

Timely Relevance Function

ln( )( )

AgbAgb i

i

tAgAgTR t e

25

Reputation Graph

• Algorithm 1: Graph Construction

26

Algorithm2: Node Evaluation

Evaluate-Node(Agy) { Arc(Agx, Agy)

If Node(Agx) is note evaluated Then Evaluate-Node(Agx)

m1 := 0, m2 := 0 Arc(Agx, Agy) {

m1 = m1 + Weight(Node(Agx)) * Weight(Arc(Agx, Agy)) m2 = m2 + Weight(Node(Agx))

} Weight(Node(Agy)) = m1 / m2

}

Algorithm 2

27

Complexity

• Construction of the trust graph with n nodes and a

edges

• n recursive calls of the function Evaluate-Node (Agy)

• Each node is visited once:

• Assessing the weight of a node

• Using the weight of its neighbors and input edges:

• Run time of the reputation algorithm:

( )O n

( )O a

(max( , ))O a n

28

Overview

Problem and Motivations

Negotiation Framework

Trustworthiness Model

• Implementation

• Application Areas

29

System Architecture

• The system is designed as a society of

interacting agents

• Agents are equipped with knowledge bases and

argumentation systems

• Knowledge bases contain propositional formulae

and arguments

• Platform: Jack Intelligent Agents + Java

30

System Architecture

31

Architecture of Negotiating Agent

32

33

34

Overview

Problem and Motivations

Negotiation Framework

Trustworthiness Model

Implementation

• Application Areas

35

Application Areas

• Web services

• E-business within semantic grid

36

Definition

• Web service (WS)

• Software application identified by a URI

• XML artifacts: Interface definition and discovering

• Web Service Description Language (WSDL)

• Universal Description, Discovery, and Integration (UDDI)

registry, ebXML

• Simple Object Access Protocol (SOAP)

37

Communities of WSs

• Gathering WSs with similar functionalities (e.g.

FlightBooking)

• Operations:

• How to initiate, set up, and specify a community of WSs?

• How to specify and manage the WSs that reside in a

community?

• How to trust WSs within a community?

38

New Architecture of WSs Communities

39

Entry Game

40

Defense Game

41

E-business in Semantic Grid

• Argumentative agents for semantic grid (ArguGrid)

42

E-business in Semantic Grid

43

Programming the Grid

Trust grid Environment

Dialogue game protocols: specification, implementation, and verification

Global View

Argumentation-based Framework for Semantic Grid

Argumentation-based Framework for Semantic Grid

Communication, Negotiation and Persuasion between Grid

Components

44

Future Work

• Evaluate the model using concrete scenarios

in e-business settings

• A general framework for secure and verifiable

grid-computing-based applications with the

underlying formal semantics

• Trust in WS communities

Argumentation and Trust: Argumentation and Trust: Issues and New ChallengesIssues and New Challenges

Jamal BentaharConcordia University (Montreal, Canada)

University of Namur, Belgium, University of Namur, Belgium, June 26, 2007June 26, 2007