Distinguished dissertations

2012 winner

For 2012, the panel chose the following winning dissertation:

Justification Based Explanation in Ontologies

Matthew Horridge, University of Manchester

Abstract:

The Web Ontology Language, OWL, is the latest standard in logic-based ontology languages. The Description Logic foundations of OWL mean that it is possible to compute what is entailed by an OWL ontology. However, without tool support, it can be very difficult or impossible to understand why an entailment holds. In the OWL world, justifications, which are minimal entailing subsets of ontologies, have emerged as the dominant form of explanation. This thesis investigates justification based explanation techniques. The core of the thesis is devoted to defining and analysing Laconic and Precise Justifications. These are fine-grained justifications whose axioms do not contain any superfluous parts. An extensive empirical evaluation shows that it is practical to compute Laconic Justifications and also reveals prevalence of non-laconic justifications in the wild. The results indicate that Laconic and Precise justifications are likely to be useful in practice.  

About the author:

Matthew Horridge completed his PhD in the Information Management Group at The University of Manchester where he was supervised by Dr Bijan Parsia and Prof Uli Sattler. Prior to this he worked as a software engineer, developing widely used APIs and tools for working with OWL ontologies. Matthew currently works in the Biomedical Informatics Research Group at Stanford University, California. His research interests centre around the conceptual, computational and cognitive aspects of explanation in ontologies.

Matthew Horridge dissertation (PDF)

Buy the dissertation from the BCS bookshop

2012 runners up

The runners up for 2012 are:

Scalable Performance Analysis of Massively Parallel Stochastic Systems

Richard Alexander Hayden, Imperial College

Abstract:

The accurate performance analysis of large-scale and distributed computer and communication systems is directly inhibited by an exponential growth in the state-space of the underlying performance model. Nevertheless, an ability to extract quantitative performance measures such as passage-time distributions from performance models of these systems is critical for service providers.

In this thesis, we develop a scalable performance analysis framework for a stochastic process algebra, facilitating the capture of key performance measures such as passage-time distributions. The approach is based on the approximation of key stochastic quantities such as means and variances by tractable systems of ordinary differential equations (ODEs). Crucially, the size of these systems of ODEs is independent of the number of interacting entities within the model, making these analysis techniques extremely scalable. The reliability of our approach is directly supported by convergence results and, in some cases, explicit error bounds.

About the author:

Richard Hayden obtained his PhD in the performance modelling group at Imperial College London, supervised by Dr Jeremy Bradley. For his undergraduate work he was awarded the 2007 BCS Computational Science Student of the Year. He is now a research associate at Imperial College where his research interests are in developing scalable asymptotic techniques for the performance analysis of massively parallel systems.

Richard Alexander Hayden dissertation (PDF)

Trust Assessment and Decision-Making in Dynamic Multi-Agent Systems

Christopher Burnett, University of Aberdeen

Abstract:

The concept of trust in multi-agent systems (MASs) has received significant attention in recent years. However, current approaches do not adequately address highly dynamic multi-agent systems, where the population and structure changes frequently.  We propose a general approach for trust evaluation and decision-making in highly dynamic multi-agent systems. First, we present a model of stereotypes, which allows agents to build tentative trust relationships with others on the basis of visible features, and propose a method of dealing with stereotypically biased reputation providers. Secondly, we present a trust decision-making model which employs controls, as well as trust evaluations and stereotypes, in order to facilitate initial interactions when trust is low or absent. We show that controls and stereotypes together can be used to bootstrap and sustain highly dynamic societies of risk-averse agents.

About the author:

Chris Burnett obtained his PhD from the University of Aberdeen, supervised by Prof. Timothy Norman and Prof. Katia Sycara (Carnegie Mellon University). Chris is currently a Research Fellow working in the dot.rural Digital Economy Research Hub at the University of Aberdeen, developing trusted mobile platforms for supporting the self-management of chronic illnesses in rural areas. His research interests lie in building effective trust and reputation mechanisms for systems comprising both human and artificial agents working together.

Christopher Burnett dissertation (PDF)