DEPARTMENT OF COMPUTING SCIENCE AND MATHEMATICS
UNIVERSITY . COMPUTING SCIENCE . SEMINARS

SEMINARS - Spring 2009

[Talk Schedule] [Abstracts] [Previous Seminars]

The Department of Computing Science and Mathematics presents the following seminars. Unless otherwise stated, seminars will take place in Room 4B94 of the Cottrell Building, University of Stirling from 15.00 to 16.00 on Friday afternoons during semester time. For instructions on how to get to the University, please look at the following routes.

If you would like to give a seminar to the department in future or if you need more information, please contact the seminar organiser, Savi Maharaj (Phone 01786 467431, Email savi@cs.stir.ac.uk).

Talk Schedule [Top] [Abstracts]

Abstracts [Top] [Schedule]

13th February [Schedule]

Goals and Policies for Telephony
Ken Turner
Department of Computing Science and Mathematics
University of Stirling

Abstract

Goals are abstract application-oriented objectives for how a system should behave. To be made operational, they are refined into lower-level policies that are executed dynamically. This talk will explain the design of ACCENT (Advanced Component Control Enhancing Network Technologies, http://www.cs.stir.ac.uk/accent) - an automated system for support of goals and policies. The approach will be explained with respect to Internet telephony, although it has also been used to manage other kinds of systems such as sensor networks and home networks.

 

6th March [Schedule]

A Symbolic Investigation of Superspreaders
Carron Shankland
Department of Computing Science and Mathematics
University of Stirling

Abstract

Superspreaders are infectious individuals who are somehow responsible for more infections in the population than average. Two main hypotheses have been presented regarding the mechanism of superspreading: supershedders and supercontacters. Supershedders transmit more disease per contact, making subsequent infection more likely, while supercontacters transmit more disease because they make more contacts in the population. Two questions
are raised: Does having superspreaders in a population affect the overall epidemiological dynamics? Does it matter which type of superspreaders are in the population?

Computing Science gives us a novel (and useful) way to look at biological phenomena. In particular, it is possible to present several hypotheses about system behaviour, to test those computationally, and to compare results with experimental data. The outcomes of this process may then be used to guide future biological experiments. In this talk I will present two very different models of superspreaders from the point of view of individual behaviour. The models are investigated by conversion to Mean Field Equations, answering the questions posed above.

 

20th March [Schedule]

Nitric Oxide diffusion: from PDE to ODE to CTMC
Andrea Degasperi
Department of Computer Science
University of Glasgow

Abstract

Our goal is to investigate whether theoretical computer science can improve the analysis of biochemical systems, as compared with traditional modelling techniques. Here we present work on a diffusion model of Nitric Oxide (NO). The end goal is a discrete state model that we can reason about by model checking. We start with a model in Partial Differential Equations (PDE), obtain an approximation in Ordinary Differential Equations (ODE) and finally develop a Continuous Time Markov Chain model that we implement in PRISM.

 

27th March [Schedule]

Cepstrum of Bispectrum - A New Approach to Bio-signal Processing
Shahjahan Shahid
Department of Computing Science and Mathematics
University of Stirling

Abstract

The human body is a source of electrical signals that provide important information to physicians and scientists about the behaviour of underlying organs including the neural system. In spite of fruitful applications of signal processing techniques in medicine, physicians and scientists are still evaluating large volumes of data relying upon manual methods: using their assumptions, experiences on data analysis, etc. One of the main reasons for this is the limitations of conventional signal processing techniques – most techniques are based on the assumptions that the observed bio-signals are linear, Gaussian and stationary whereas, in practical, biomedical signals are complex structures and hence violate all the above assumptions.

Many biological signals can be modelled as the filtered impulse of time series. Higher order statistical (HOS) methods are useful for analyzing this type of stochastic process. HOS methods have the advantage of analyzing any non-Gaussian and non linear signal. We developed a new HOS based statistical measurements - Cepstrum of Bispectrum (CoB). It carries all the basic properties of HOS and has a few new features. In this seminar we will observe some properties of HOS and CoB and its applications to some real world bio-signals such as ECG, EMG and EEG.

 

 

3rd April [Schedule]

Optimising paired and pooled kidney exchanges
David Manlove
Department of Computer Science
University of Glasgow

Abstract

Recent changes in legislation in the UK have allowed a patient who requires a kidney transplant, and who has a willing but incompatible donor, to be involved in a "pairwise kidney exchange" with another patient-donor pair in a similar position. "Pooled donations" extend this concept to three couples in a cyclic manner. UK Transplant run the National Matching Scheme for Paired Donation, which finds pairwise and 3-way exchanges (the latter being pooled exchanges involving three couples) at regular intervals. This matching scheme is based on optimising firstly the number of transplants, and subject to this, the total weight of the transplants, based on a scoring system that incorporates a number of factors including sensitivity, HLA compatibility, age and geographic location.

In this talk we describe a range of computational techniques that we have used in order to construct optimal exchanges for UK Transplant on various occasions between April 2008 and January 2009. The first two of these involve polynomial-time algorithms, based on weighted matching in graphs, to find an optimal set of exchanges and are implemented in C++. The third technique deals with an NP-hard optimisation problem, and uses integer linear programming, implemented in Matlab. We present some computational results based on simulations involving random samples from a large dataset of patient-donor pairs, and in addition we present a summary of the results that we obtained for the "real" matching runs, as applied to anonymous data supplied by UK Transplant.

This is joint work with Péter Biró and Kirstin Macdonald.

 

 

3rd April [Schedule]

Firefly: Highlighting future trends in computer systems
Joe Finney
University of Lancaster

Abstract

Computers have become commodity items. Most people now treat the presence of a computer in the home or at work with about the same enthusiasm as a washing machine, and they cost about the same amount too. In fact, home computers can now often be found on sale in supermarkets somewhere between the cheese crackers and last year’s Easter eggs...

Some people believe the commoditisation of computing spells the end of the recent boom in the development and utilisation of innovative computer systems and applications. In reality, this couldn’t be further from the truth. Computing is being applied to an increasingly diverse range of new and exciting scenarios, and this is resulting in the development of new and novel technologies, many of which we use every day without giving them a second thought. We are quietly slipping into a whole new era of computer systems.

Project Firefly is an ongoing research and development project within the Computing department at Lancaster that is investigating ways of building coherent, self-organizing display surfaces. In other words, bringing the pixels off a computer screen and into the real world.

Using Firefly as a case study, this talk will provide a conceptual overview of how the technology operates, the advantages it brings, and through this will highlight some future trends, challenges and opportunities in computing.

 

24th April [Schedule]

Automated Software Verification - A Renaissance
Dependable Systems Group
School of Mathematical and Computer Sciences
Heriot-Watt University

Abstract

The goal of developing software that can be formally verified as correct with respect to its intended behaviour has a long and distinguished history, dating back over fifty years. Much has been achieved, but what has remained elusive are scalable verification tools that can deal with the complexities of software systems. However, times are changing, as reflected in what could be called a renaissance within the formal software verification community. My talk will focus on automated reasoning aspects of software verification. In particular, I will argue for the benefits that can be achieved if complementary reasoning techniques are combined cooperatively. As evidence for my argument I will draw upon a past industrial oriented project. Some new research projects will also be highlighted together with research challenges.

1st May [Schedule]

Developing a Framework for the Multiverse
Karla Parussel
Department of Computing Science and Mathematics
University of Stirling

Abstract

The Interlife project seeks to use virtual environments such as Second Life and Open Sim to aid young people going through difficult social transitions. In this talk I discuss how we are developing a framework to handle communications between these environments and devices in real life. I also talk about how objects in these virtual environments can be scripted to communicate with each other. Examples discussed include a surveillance
grid, a diary room, a percussion instrument and an exploding flatulent rocket.

 

8h May [Schedule]

Predictive Detection of Feature Interaction
Rui Crespo
Technical University of Lisbon

Abstract

Internet applications, such as Email, VoIP and WWW, have been enhanced with many features. However, the introduction and modification of features may result in undesired behaviours, and this effect is known as feature interaction-FI.

We propose a proactive approach for FI detection. Supported by sets of all possible events, predicates and inconsistent behaviours, we generate hypothetical new features that interact with a given feature. By predicting FIs, the feature subscriber may define, in advance, all mechanisms to resolve the FIs that may occur in the future.

We adopt a semantic model, based on group theory, for the feature axiomatic specification. The algorithms that generate new features do not depend on the particular data structures used in the semantic model of the feature specification.

Previous Seminar Series [Top] [Abstracts] [Schedule]

Last Modified: 25th January 2008