Back to the Future: an Interaction-oriented Framework for Social - - PowerPoint PPT Presentation

Requirements Engineering for Social Computing @ RE 2011

Back to the Future: an Interaction-oriented Framework for Social Computing

• M. Baldoni1
• C. Baroglio1
• E. Marengo1
• V. Patti1
• A. Ricci2

1Dipartimento di Informatica, Universit`

a degli Studi di Torino

2DEIS, Universit`

a degli Studi di Bologna

Trento, August 29, 2011

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 1 / 19

Overview

1

Vision and motivation

2

An interaction-centric framework

3

Discussion and conclusions

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 2 / 19

B2B, cross-business, open environment systems

Facilitate development of new products Cooperatively exploit resources Share best practices Integration Interaction based on agreed contracts

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 3 / 19

B2B, cross-business, open environment systems

Usually, the integration is based on the classical notion

• f control flow of their

software, even by means of

• rchestration languages

However, this does not help software reuse and modular development Instead, this reality demands abstraction and models where the involved entities are fully autonomous

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 4 / 19

An interaction-centric approach

What is it necessary?

A notion of “coordination”, obtained by introducing social dimension A new interaction-centric approach Interaction, coordination, and communication are all central issues to the area of MAS but the current platforms are too much content-centric

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 5 / 19

A new equation?

Some former equations

Algorithms + Data Structures = Programs Algorithm = Logic + Control

The new equation

Social computing = Dependencies + Autonomous Control

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 6 / 19

A new equation?

Some former equations

Algorithms + Data Structures = Programs Algorithm = Logic + Control

The new equation

Social computing = Dependencies + Autonomous Control

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 6 / 19

Normative characterization of interaction

Interaction create social expectations and bindings/dependencies, but ... ... a normative characterization of coordination is needed [Castelfranchi, 1997, Singh, 1999], so that ... ... the publicly acceptance

• f the regulation allows

reasoning about agents’ behavior [Conte et al., 1998]

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 7 / 19

Systems made of autonomous and heterogeneous components

Current platforms do not supply: agents the means for

• bserving or reasoning

about such meanings of interaction the designers the means to explicitly express and characterize them when developing an interaction model

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 8 / 19

Agents and Artifacts as abstractions

A&A meta-model [Weyns et al., 2007, Omicini et al., 2008] provides abstractions for environments and artifacts, that can be acted upon, observed, perceived, notified, . . . From a SE point of view:

◮ Abstraction ◮ Modularity and

encapsulation

◮ Extensibility and

adaptation

◮ Reusability BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 9 / 19

(In)Direct Communication

Environments, artifacts can be perceived, acted upon, observed, . . . All interactions among agents will be indirect, like in the real world As Keil and Goldin observed, indirect communication fosters the collaboration and the coordination inside open systems Environments, artifacts can be general, programmable channels of communication

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 10 / 19

Agents and Artifacts as abstractions

Why do we use enviroments and artifacts?

To reify regulations aimed at coordination Agents can examine them Agents can use them Agents can construct them

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 11 / 19

Agents and Artifacts as abstractions

Introducing a normative characterization

Commitment-based approach [Castelfranchi, 1997, Singh, 1999] A semantics of interaction for design time verification Artifacts and environments for runtime verification Social computing = Dependencies + Autonomous Control

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 12 / 19

Agents and Artifacts as abstractions

Introducing a normative characterization

Commitment-based approach [Castelfranchi, 1997, Singh, 1999] A semantics of interaction for design time verification Artifacts and environments for runtime verification Social computing = Dependencies + Autonomous Control

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 12 / 19

The schema of our proposal

The specification level allows the designer to shape the interactions that will characterize the system by supplying adequate high-level abstractions The programming abstraction level realizes at a programming language level the abstractions defined above Our starting point for the infrastructure level is the CArtAgO framework [Ricci et al., 2009]

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 13 / 19

Specification Level

We propose to rely upon commitment-based approach and, in particular, commitment-based protocols [Singh, 1999, Singh, 2000, Yolum and Singh, 2001] Standardization and regulation of interaction is a decisive factor in distributed and open systems, made of heterogeneous and changing parties We rely on the proposal in [Baldoni et al., 2011, Marengo et al., 2011], that allows the representation of legal patterns of interaction by enriching commitment protocols with temporal regulations.

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 14 / 19

Specification Level

In particular, [Baldoni et al., 2011] proposes a decoupled approach that separates a constitutive and a regulative specification The constitutive specification defines the meaning of actions based on their effects on the social state The regulative specification reinforces the regulative nature of commitment by adding a set of behavioral rules, by means of temporal constraints among commitments Defining the legal evolution of the social state, independently from the executed actions Advantages: easier re-use of actions, easier customization, greater compositionality

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 15 / 19

Programming Abstractions Level

Incorporating interaction protocols based on commitments, patterns

• f interaction, forms of direct and indirect communication and

coordination between agents (such as stigmergic coordination) inside the programmable environments envisaged by the A&A meta-model [Weyns et al., 2007, Omicini et al., 2008] The act of using an artifact can be interpreted as a declaration of acceptance of the coordination rules This will generate social expectations about the agent’s behavior and this agrees with the characterization of norms in [Conte et al., 1998]

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 16 / 19

Programming Abstractions Level

Incorporating interaction protocols based on commitments, patterns

• f interaction, forms of direct and indirect communication and

coordination between agents (such as stigmergic coordination) inside the programmable environments envisaged by the A&A meta-model [Weyns et al., 2007, Omicini et al., 2008] The act of using an artifact can be interpreted as a declaration of acceptance of the coordination rules This will generate social expectations about the agent’s behavior and this agrees with the characterization of norms in [Conte et al., 1998]

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 16 / 19

Programming Abstractions Level

Commitment-based protocols encoded into artifacts Programmable communication media, having a normative characterization

Software Engineering point of view

This conjugation helps providing protocol specifications good and important features: abstraction, modularity and encapsulation, extensibility and adaptation, reusability

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 17 / 19

Infrastructure Level

Environments used as a computational support for the agents’ activities [Omicini et al., 2004] Artifacts used for realizing stigmergic coordination mechanisms [Ricci et al., 2007], organizational artifacts [Hubner et al., 2009, Piunti et al., 2009] Our starting point is the CArtAgO framework [Ricci et al., 2009]

◮ a proper computational model, and ◮ a programming model for the design and the development of the

environments on the base of the A&A meta-model

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 18 / 19

Conclusion

Interaction-centric Social meaning of interaction Programmable Communication channel with monitoring functionalities A normative value thanks to commitmentbased approach Explicit acceptance of the regulamentations Flexibility and openess typical of MAS Modularity and compositionality typical of design and development methodologies

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 19 / 19

Baldoni, M., Baroglio, C., Marengo, E., and Patti, V. (2011). Constitutive and Regulative Specifications of Commitment Protocols: a Decoupled Approach. ACM Trans. on Int. Sys. and Tech., Spec. Iss. on Agent Communication. To appear. Castelfranchi, C. (1997). Principles of Individual Social Action. In Holmstrom-Hintikka, G. and Tuomela, R., editors, Contemporary action theory: Social action, volume 2, pages 163–192, Dordrecht. Kluwer. Conte, R., Castelfranchi, C., and Dignum, F. (1998). Autonomous Norm Acceptance. In Proc. of ATAL, volume 1555 of LNCS, pages 99–112. Springer. Hubner, J. F., Boissier, O., Kitio, R., and Ricci, A. (2009). Instrumenting multi-agent organisations with organisational artifacts and agents: “Giving the organisational power back to the agents”.

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 19 / 19

Autonomous Agents and Multi-Agent Systems, 20. Marengo, E., Baldoni, M., Baroglio, C., Chopra, A. K., Patti, V., and Singh, M. P. (2011). Commitments with Regulations: Reasoning about Safety and Control in REGULA. In Proc. of AAMAS. Omicini, A., Ricci, A., and Viroli, M. (2008). Artifacts in the A&A meta-model for multi-agent systems. Autonomous Agents and Multi-Agent Systems, 17(3):432–456. Omicini, A., Ricci, A., Viroli, M., Castelfranchi, C., and Tummolini, L. (2004). Coordination Artifacts: Environment-Based Coordination for Intelligent Agents. In 3rd International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS 2004), pages 286–293, New York, USA. Piunti, M., Ricci, A., Boissier, O., and H¨ ubner, J. F. (2009). Embodying Organisations in Multi-agent Work Environments.

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 19 / 19

In Proceedings of the 2009 IEEE/WIC/ACM International Conference

• n Intelligent Agent Technology, IAT, pages 511–518. IEEE.

Ricci, A., Omicini, A., Viroli, M., Gardelli, L., and Oliva, E. (2007). Cognitive stigmergy: Towards a framework based on agents and artifacts. In Proc. of E4MAS, volume 4389 of LNCS, pages 124–140. Springer. Ricci, A., Piunti, M., Viroli, M., and Omicini, A. (2009). Environment Programming in CArtAgO. In Multi-Agent Programming II: Languages, Platforms and Applications, Multiagent Systems, Artificial Societies, and Simulated Organizations. Singh, M. P. (1999). An Ontology for Commitments in Multiagent Systems.

• Artif. Intell. Law, 7(1):97–113.

Singh, M. P. (2000). A social semantics for agent communication languages.

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 19 / 19

In Issues in Agent Communication, volume 1916 of LNCS, pages 31–45. Springer. Weyns, D., Omicini, A., and Odell, J. (2007). Environment as a first class abstraction in multiagent systems. Autonomous Agents and Multi-Agent Systems, 14(1):5–30. Yolum, P. and Singh, M. P. (2001). Commitment machines. In Proc. of ATAL, pages 235–247.

BBMPR (UniTO&UniBO) RESC 2011 Trento, August 29, 2011 19 / 19