Project

Objectives of the project:

Over the last 25 years one of the most important architectures for autonomic software agent systems has been the Belief-Desire-Intention (BDI) architecture which came originally from the work of Bratman and was developed in many institutions most notably during the time of the Australian Artificial Intelligence Institute under the directorship of Michael Georgeff. The first computational model that gave a formal and comprehensive account of the activity of such agents was developed by a team led by d'Inverno (one of the co-investigators) in the early 90s to provide a systematic attempt to describe the operation of such agents that became the foundation for various instantiations to be built. BDI agents have many of the
properties of autonomic agents as discussed in the call in that (a) they maintain a symbolic representation of their current beliefs, goals, plans and intentions, (b) are aware of their dynamically changing environment within it (c) model others (d) dynamically respond to changes in their environments (e) make predictions about the intended outcome of actions (f) argue and negotiate with other agents and can modify their models of each other and the humans users they represent.
 
Whilst there has been a huge amount of work in developing infrastructures for agents to interact and coordinate their activities, nearly all the work that has been presented (workflows, business models, Petri nets) actually inhibits the autonomy and reflective decision making of agents. Perhaps the most important counterexample is the work of
electronic institutions that provides the ability for autonomic agents to interact in a truly autonomous way within the society of a multi-agent context. Over the last 15 years this work has been developed by IIIA under the leadership of Carles Sierra, one of the named investigators for this project. Essentially an electronic institution can be thought of as a set of interconnected scenes such as we might find in a theatre. In these scenes agents take on roles and interact with each other in order to achieve their individual goals. Once an agent leaves a scene it can then choose which next scene it would like to be involved with and with what role. Thus the system provides a non-centralised architecture that facilitates meaningful interaction without impacting on the autonomy of individual agents.
 
We believe centralised solutions to enabling the development of properly autonomic societies of agents do not work. We must go to a peer to peer structure where people can freely leave and join such systems and to this end we will develop an open system that enables human agents to join and leave and that encourages a deeper shared and social cultural experience than is now currently possible.
 
Decades of cultural institutional digitisation projects and individual website development have led to a large body of European culture available in digital archives and the internet: the material includes digital reproductions of physical (visual, aural, and video) works and writing, speaking, and video analysis and responses to original artworks. And of course this body is growing rapidly. One of the goals of ACE is to make this material more useful by making it more engaging. The sense of engagement will come from enabling the sites to become collective spaces in which users will be able to share their views of this material and discuss the material and engage with the materials as groups.
 
This project is therefore not about cultural or knowledge content but about the distributed and social consumption and interaction with cultural artefacts. In short, we believe BDI architectures represent the strongest and best-developed software engineering device for building autonomic agents and that electronic institutions is the best developed infrastructure for supporting the interaction of autonomous interaction.
 
This project is about exploiting the predominance of social networking to enrich, encourage and enliven engagement with online cultural artefacts such as from a museum or a gallery. With the current problems in the European financial debt many cultural institutions are planning to shorten the length that visitors can physically enter. In the UK for example we have heard or plans that the British Museum will close earlier and possibly shut down completely for one day a week because of the massive cuts in funding that were presented in the UK Chancellors speech detailing reduction in money for the cultural sector. Coupled with this problem is the ever present and perhaps increasing desire to enable more young people to engage with our cultural institutions to not only enrich their own lives but also, in turn, enrich the culture of our European Countries themselves. The predominance of social networking in young people lives has surprised us all. And many of us are aware of how our teenage children whilst watching the television are in constant contact with their friends describing and discussing the various fictional and dramatic scenarios as they unfold before
them.
 
We will harness the power of autonomic agents that work on behalf of human users in an infrastructure that allows for these agents to communicate and negotiate on behalf of their human users to facilitate a collective and social experience of online cultural visits. For example, we could imagine a scenario where 4 students are visiting an art museum with the desire to purchase something (a print or a physical copy of an artefact for example) for a friend. They would wish to be able to negotiate about what to see or experience online, what additional information they want to consider, what comments from what previous visitors over any commentary they individually or collectively want to leave for others and, eventually, over what they collectively choose to purchase for their friend. We believe this context provides a timely and important case study for developing autonomic agent architectures. Specifically, in order to facilitate our project, we are concerned with the fundamental question of building autonomic agents that can represent their users needs, maintain models of other autonomic agents and proactively develop plans and solutions for their human counterparts. In order to achieve this we will develop this architecture based on our combined expertise in BDI agents – autonomic agents that symbolically represent their belief, desires and intentions and those of other agents. Such agents are aware of their environment and others and plan to act in the world, possibly collaboratively, in order to satisfy their goals.
 
The overall objective of this project is to demonstrate the added value of using autonomic agents to enrich and enable the social experience that is potentially available when visiting online cultural institutions and lay the foundation for building increasingly sophisticated agent and interaction architectures for enabling critical cultural activity.
 
The specific aims of this project are:
  • Develop autonomic BDI architectures for the personal assistants of human agents engaging in online activity
  • Develop a peer to peer autonomic electronic institution infrastructure to support the autonomous interaction or human and autonomic agents
  • Embed the P2P infrastructure into mobile appliances to allow for a mobile social distributed consumption of cultural artefacts
  • Develop a series of case studies with cultural institutions and build a full specification of a selected case study that we will develop into a working prototype
  • Demonstrate the value of truly autonomous agents in building systems that expand the possibilities for socially engaging with online activity.
  • Develop formal models of the agent architectures and underlying electronic institution architecture. This is important because we do not want the work to be "lost" as soon as the project stops but to be the foundation for future work in the development of autonomic agents and infrastructure to properly support autonomy.
  • Similar in spirit to 6, provide the formal and computational architecture (as open source freeware) for the further development of the research sector
  • Develop and publish a self-evaluation document on working with the cultural sector using new technology innovations around autonomic personal agents.