David Bonnefoy

How to Get the Recommender Out of the Lab

A personalised system is a complex system made of many interacting parts, from data ingestion to presenting the results to the users. A plethora of methods, tools, algorithms and approaches exist for each piece of such a system: many data and metadata processing methods, many user models, many filtering techniques, many accuracy metrics, many personalisation levels. In addition, a realworld recommender is a piece of an even larger and more complex environment on which there is little control: often the recommender is part of a larger application introducing constraints for the design of the recommender, e.g. the data may not be in a suitable format, or the environment may impose some architectural or privacy constraints. This can make the task of building such a recommender system daunting, and it is easy to make errors. Based on the experience of the authors and the study of other works, this chapter intends to be a guide on the design, implementation and evaluation of personalised systems. It presents the different aspects that must be studied before the design is even started, and how to avoid pitfalls, in a hands-on approach. The chapter presents the main factors to take into account to design a recommender system, and illustrates them through case studies of existing systems to help navigate in the many and complex choices that have to be faced.

More Like This or Not for Me: Delivering Personalised Recommendations in Multi-user Environments

The television as a multi-user device presents some specificities with respect to personalisation. Recommendations should be provided both per-viewers as well as for a group. Recognising the inadequacy of traditional user modelling techniques with the constraint of televisions lazy watching usage patterns, this paper presents a new recommendation mechanism based on anonymous user preferences and dynamic filtering of recommendations. Results from an initial user study indicate this mechanism was able to provide content recommendations to individual users within a multi-user environment with a high level of user satisfaction and without the need for user authentication or individual preference profile creation.

Dealing with interoperability for agent-based services

Interoperability problems occur when developing agent systems with a strong notion of autonomy while interacting with other agents. The agent platform has become the intelligent distributed operating system for agent applications and therefore the support of much of the interoperability between agents will be through the platforms themselves. With the possibility of inter- platform interoperability comes support for open services that goes beyond current open agent architectures. This paper looks at the design issues of interoperability that need to be addressed on both the communication and application levels. It identifies solutions to some of these issues and defines a set of components, which are implemented and tested as a toolkit, in order to support Open Agent Service Architecture development and deployment.

Agent Based Dynamic Service Synthesis in Large-Scale Open Environments: Experiences from the Agentcities Testbed

The notion of autonomous agents populating large-scale open environments, such as the public Internet, that are able to dynamically discover one another, interact and synthesise new software applications or results has become one of the key technology visions of the past few years. This Agentcities testbed represents one of the largest attempts to date to prototype such a vision: deploying current generations of agent and Semantic Web technologies to create a global test bed for dynamic service composition involving more than 100 participating organisations. The paper presents an overview of this initiative.