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

SEMINARS - SPRING 2004

[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 organisers, either David Cairns (Phone 01786 467445, Email dec@cs.stir.ac.uk) or Julie Cowie (Phone 01786 467446, Email jco@cs.stir.ac.uk).

Talk Schedule [Top] [Abstracts]

Abstracts [Top] [Schedule]

20th February [Schedule]

Learning Internal Models For Control
Dr Sethu Vijayakumar
School of Informatics, University of Edinburgh

Abstract: Humans and other Biological Systems are very adept at performing fast, complicated control tasks inspite of large sensorimotor delays while being fairly robust to perturbations. The key to this is the ability to learn "internal models" -- which from a control theoretic point of view are nothing but highly nonlinear forward (or inverse) dynamics models. In this talk, I will present a framework for real-time, online learning of internal model for control and demonstrate it's applicability to the control of high dimensional, compliant humanoid robots. I will also present a novel dynamical system based trajectory planning mechanism which can be used to prime the learning of these models very efficiently.

This seminar is jointly organized with the EPSRC Novel Computation Cluster.

 

27th February [Schedule]

Model checking and planning. How similar are they?
Dr Alice Miller
Department of Computing Science, University of Glasgow

Some planning problems can be expressed as logical properties that can be checked using model checking. Recently there has been increased interest in the problem of using Symmetry breaking during planning. Symmetry reduction techniques are fairly standard in model checking, although their implementation tends to be somewhat limited thanks to the well-known "orbit problem". In this talk I will investigate the relationship between planning and model checking. I will give an example of a classic planning problem and show how it can be solved using the SPIN model checker. I shall also survey recent symmetry-related results and discuss how the two communities could benefit each other.

 

5th March [Schedule]

Sound Feature Detection Using Leaky Integrate-And-Fire Neurons
Dagmar Fraser
Department of Computing Science, University of Stirling

We present a neurally inspired technique for detecting onsets and amplitude modulation in sound. This starts with a cochlea-like filter. The outputs from this filter are spike coded, in a way similar to the auditory nerve (AN). These AN-like spikes are presented to leaky integrate-and-fire (LIF) neurons through a depressing synapse. The spike outputs from these are then processed by another layer of LIF neurons. Onsets are detected with essentially zero latency. Amplitude modulation is detected in a way similar to that of onset chopper cells. We present results from some of the TIMIT database.

12th March [Schedule]

OSGi, Networked Devices and Services within the home context
Michael Wilson
Department of Computing Science and Mathematics, University of Stirling

Home networking and automation is an area that is enjoying considerable growth with new protocols, devices and services being developed for the smart home. At present home networking is limited to a shared Internet connection, printer or some other computer peripheral. In the not so distant future this sort of networking will be broadened out to include everyday items such as a toaster, lamp, heater, sensor, televisions and so on. In the home the OSGi (Open Service Gateway Initiative) platform hosts home-services that control home devices. OSGi can be seen as the glue which allows services to manipulate any device, regardless of its underlying protocol. This allows services to become more flexible using whatever devices are available to them. This talk will look at some of the current home networking protocols and devices and discusses how OSGi is used as the glue which allows services to become more powerful, fulfilling the potential of consumer devices.

 

26th March[Schedule]

Grids & Life Sciences
Dr Richard Sinnott
National e-Science Centre, University of Glasgow

The Grid is an infrastructure to enable collaborative research through so called virtual organisations. Typical artefacts of virtual organisations are the need to access and use large computational resources and to deal with potentially vast and distributed data sets. Scientific research applying Grid technology is typically referred to as e-Science. The life sciences represent an especially challenging area for Grid technologies for several reasons. Firstly the data sets associated with life sciences are growing exponentially, are distributed and cover a wide variety of areas including amongst other things nucleotide sequences, proteins and protein structures, cells and cell signalling right through to whole organisms. Secondly these data sets are often dirty with erroneous and contradictory data which itself may have evolving schemata and potentially different security requirements.

In this talk we will provide an overview of life science issues in applying Grid technologies. An assessment of where Grid technologies are today will be presented as well as where the technologies are leading us tomorrow with emphasis upon in silico life science research. The role of the National e-Science Centre in organising and coordinating national level e-Science initiatives will also be outlined.

 

About The Speaker
Dr Sinnott has been Technical Director of the National e-Science Centre at the University of Glasgow since December 2002. In addition to this, Dr Sinnott has the role of Deputy Director (Technical) of the Bioinformatics Research Centre headed by Prof. David Gilbert. Prior to coming to Glasgow, Dr Sinnott ran his own consultancy company based in Germany specialising in the area of formal technologies and their application to real time systems development, especially in the telecommunications domain. Before this he worked as a senior research scientist at GMD Fokus in Berlin, where he worked on and managed various international projects in the telecoms, service management and network management domains. He holds a PhD from the University of Stirling where his research was based on the modelling and architectural design of open distributed processing systems. He edited several international standards in this domain. He also holds an MSc in Software Engineering from the University of Stirling (dissertation on the Formal Specification of Electronic Components in LOTOS) and a BSc in Theoretical Physics from the University of East Anglia (UEA) in Norwich (dissertation on Computer Simulation of a Linear Polymer).

Dr Sinnott's current research is focused around Grid computing using technologies such as the Globus Toolkit and its application to a broad spectrum of scientific areas - bioinformatics being one exemplar. Dr Sinnott also maintains an interest in formal methods and their application to real time, distributed systems development.

 

2nd April [Schedule]

Evolving better cancer therapies using genetic algorithms
Dr. John McCall
School of Computing, The Robert Gordon University

Cancer is a major disease in Western Europe and North America, accounting for one in three of all deaths. The problem is particularly acute in Scotland where, for a range of reasons, both the incidence of cancer and mortality rates from many forms of the disease are higher than in other comparable nations.

This talk describes how genetic algorithms (GAs) can assist in the design of more effective cancer treatments. I will describe approaches to modelling tumour growth and the effect of cancer chemotherapy in restricting growth. GAs can use such models to try to evolve a best-possible feasible drug-delivery schedule that meets certain objectives, singly or as part of a multi-objective optimisation.

Typical objectives include tumour eradication, minimising the severity of toxic side effects, and prolonging the lives of patients with incurable cancers. I will also describe a medical decision support system that allows clinicians to use the GAs and tumour simulations in an interactive way.

 

16th April [Schedule]

The Tao of PEPA nets
Stephen Gilmore
School of Informatics, University of Edinburgh

PEPA nets are an amalgamation of stochastic process algebras and stochastic Petri nets which are used for modelling the long-run and transient behaviour of systems which are composed of concurrently active stateful mobile objects. This talk will introduce the PEPA nets language and provide examples of its use.

 

23rd April [Schedule]

Software Rollout in (Tele)communications Networks : Problems and Processes
Mussy Kurt Elli
edNET

There are many areas of concern in telecommunications systems during enhancements or extensions of networks. The nature of the service, being real time delivery of payloads (speech, video or data), is a major factor in the design as well as the methodology of software rollout. The sheer size, network complexity and the distributed nature causes interesting problems. In this presentation, I will address the requirements, problems, design approach and rollout approaches, not only for the software but also for the data held within the systems. In this presentation I will include the experience of edNET in its journey to becoming a telecoms operator.

 

30th April [Schedule]

Black Box Checking
Professor Doron Peled
Department of Computer Science, University of Warwick

Two main approaches are used for increasing the quality of systems: in model checking, one checks properties of a known design of a system; in testing, one usually checks whether a given implementation, whose internal structure is often unknown, conforms with an abstract design. We are interested in the combination of these techniques. Namely, we would like to be able to test whether an implementation with unknown structure satisfies some given properties. We propose and formalize this problem of black box checking and suggest several algorithms. Since the input to black box checking is not given initially, as is the case in the classical model of computation, but is learned through experiments, we propose a computational model based on games with incomplete information. We use this model to analyze the complexity of the problem. We also address the more practical question of finding an approach that can detect errors in the implementation before completing an exhaustive search.

 

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

Last Modified: 6th March 2003