This page contains information about major current and previous research projects. Some of our projects are funded by external research funding organisations. Please refer to individual members' home pages for more details on their activities.

STUDYING THE EVOLUTION OF A NETWORK OF LEARNING, SELF-ORGANISING AGENTS (ongoing)

Distributed systems are becoming more and more popular in today's world because of their reliability and ease of deployment. To cope with large scale, these systems are typically organized in a network. The Internet and peer-2-peer are just few examples of such systems. One solution allows agents to learn how to work together and at the same time restructure the network in order to optimize performance (Abdallah and Lesser 2007). The paper verifies experimentally the effectiveness of the solution but for a specific problem (task allocation), but it would be better to test the idea in more generic context. We are currently pursuing several projects in that area: Developing and evaluating a generic framework for multiagent learning and self-organization in networks. Using data mining techniques on the sequence of networks to discover interesting patterns and behaviours. This includes: implementing a visualization tool that displays the network evolution overtime, extracting network features (e.g. diameter, average node degrees, etc.) and performing data mining techniques over these features.

Personnel:

• Sherief Abdallah
• Rana Aref (Research Assistant).

Collaborators:

• Victor Lesser (University of Massachusetts, Amherst, USA).
• Chongjie Zhang (University of Massachusetts, Amherst, USA).
• Muaz Niazi (Foundation University, Islamabad, Pakistan).

Sample Publications:

• S. Abdallah and V. Lesser. Multiagent Reinforcement Learning and Self-organization in a Network of Agents. In Proceedings of 6th International Joint Conference on Autonomous Agents and Multi-agent Systems, AAMAS, 2007 (acceptance rate 22%) runner-up for "Best Paper Award
• S. Abdallah (2010). Using Graph Analysis to Study Networks of Adaptive Agent. In: Proceedings of 9th International Conference on Autonomous Agents and Multiagent Systems (AAMAS)
  

ARGUMENTATION AND GAME THEORY (2007-)

This project investigates the use of game theory to analyse argumentation in multi-agent systems. The work explores two aspects of argumentation: (1) the design of the interaction mechanism (i.e. rules of dialogue) to ensure agents have incentive to argue in a certain way; and (2) the design of adaptive agent argumentation strategies. This work produced a mapping of Dung's abstract model of argument as a Mechanism Design problem. This enables the use of tools from game theory for analysing strategic interaction among argumentative agents.

Personnel:

• Iyad Rahwan
• Sherief Abdallah

Collaborators:

• Kate Larson (University of Waterloo, Canada).

Sample Publications:

• I. Rahwan, K. Larson and F. Tohmé (2009). A Characterisation of Strategy-Proofness for Grounded Argumentation Semantics. In: Proceedings of the 21st International Joint Conference on Artificial Intelligence (IJCAI), Pasadena, California, USA.
• I. Rahwan and K. Larson (2008). Mechanism Design for Abstract Argumentation. In: Proceedings of 7th International Conference on Autonomous Agents and Multiagent Systems (AAMAS). Estoril, Portugal.
• I. Rahwan and K. Larson (2008). Pareto Optimality in Abstract Argumentation. In: Proceedings of 23rd Conference on Artificial Intelligence (AAAI). AAAI Press, California, USA. (Acceptance rate 24%)

ARGUMENTATION AND THE SEMANTIC WEB (2006-2009)

This project investigates the use of Semantic Web languages and tools to develop systems for supporting argumentation on the Web. We design ontologies (e.g. using RDF and OWL) for describing arguments and we build prototype systems that use these ontologies to enable users to navigate and manipulate complex argument structures. The work builds in the recently proposed Argument Interchange Format (AIF).

Personnel:

• Iyad Rahwan
• Sohan D'Souza (Research Assistant)
• Bita Banihashemi (past MSc student).
• Fouad Zablith, Open University, UK (past MSc student).

Collaborators:

• Chris Reed (University of Dundee, UK)

Sample publications:

• I. Rahwan, F. Zablith and C. Reed. Laying the Foundations for a World Wide Argument Web. Artificial Intelligence, Vol 21, No 6, pages 489-527
• I. Rahwan, F. Zablith, and C. Reed (2007). Towards Large Scale Argumentation Support on the Semantic Web. In: Proceedings of 22nd Conference on Artificial Intelligence (AAAI). AAAI Press, California, USA..
• C. Chesñevar, J. McGinnis, S. Modgil, I. Rahwan, C. Reed, G. Simari, M. South, G. Vreeswijk and S. Willmott (2006). Towards an Argument Interchange Format. The Knowledge Engineering Review, Vol 21, No 4, pages 293-316.

INTEREST-BASED NEGOTIATION: THEORY AND PRACTICE (2004-2007)

The project, funded by the Australian Research Council, builds on prior work from Iyad Rahwan's PhD concerning a novel form of automated negotiation - interest-based negotiation - where participants (software agents) exchange arguments about their goals and desires during a structured dialogue. The primary focus of this project is to refine the negotiation framework and to evaluate its theoretical properties. View project home page here.

Personnel:

• Iyad Rahwan (co-investigator)

Collaborators:

• Liz Sonenberg, University of Melbourne (principal-investigator)
• Frank Dignum, Utrecht University (co-investigator)
• Philippe Pasquier (post-doctoral fellow)

Sample publications:

• I. Rahwan, P. Pasquier, L. Sonenberg, and F. Dignum (2007). On the Benefits of Exploiting Underlying Goals in Argument-based Negotiation. In: Proceedings of 22nd Conference on Artificial Intelligence (AAAI). AAAI Press, California, USA.
• P. Pasquier, L. Sonenberg, I. Rahwan, F. Dignum, R. Hollands (2007). An Empirical Study of Interest-based Negotiation. In: Proceedings of 9th International Conference on Electronic Commerce (ICEC). ACM Press, USA (Winner of the Best Paper Award)
  

AUTOMATED INTERACTION TECHNOLOGIES FOR MOBILE HANDL-HELD DEVICES (2004-2006)

This project, funded by the Australian Research Council, is concerned with developing intelligent technologies that support human interaction through handheld devices, connected to wireless digital networks. Intelligent software assistants running on these devices will interact with one another in order to coordinate the activities of multiple users. Such automated coordination may facilitate scheduling of meetings, collaboration among different members of a team, and negotiating conflicts that may arise between the activities of different connected users. This will enable context-dependent task support and coordination of workers involved in industrial applications involving mobility and dynamism.

Personnel:

• Iyad Rahwan (co-investigator)

Collaborators:

• Liz Sonenberg, University of Melbourne (principal-investigator)
• Frank Dignum, Utrecht University (co-investigator)

Sample publications:

• I. Rahwan, F. Koch, C. Graham, J. A. Kattan and L. Sonenberg (2005). Goal-directed Automated Negotiation for Supporting Mobile User Coordination. In A. Dey, B. Kokonov, D. Leake and R. Turner (Eds.) Modeling and Using Context: 5th International and Interdisciplinary Conference, CONTEXT 2005, Paris, France, July 2005, Lecture Notes in Artificial Intelligence, Vol 3554, Springer-Verlag, Berlin, Germany.