Skip to main page content - your browser does not fully support our CSS, or is text-only.

Computing Science Seminars, 2017/2018

Spring 16 image

Seminars will take place in Room 4B96,  Cottrell Building, University of Stirling. Normally, from 15.00 to 16.00 on Friday afternoons during semester time, unless otherwise stated. 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, Dr. Jess Enright (jae@cs.stir.ac.uk)

Autumn 2017

Date Speaker Title/Abstract
Friday
8th Sept
Dr. Dalila Hamami Controlling and understanding infectious diseases through data mining and modeling   Abstract: Vaccination programs for childhood diseases, such as measles, mumps and rubella have greatly contributed to decreasing the incidence and impact of those diseases. Nonetheless, despite long vaccination programmes across the world, mumps has not yet been eradicated in those countries. A resurgence of mumps disease has been investigated by a massive number of computational models to assist decision-making in public health epidemiology. However, achieving the best model is a complex task due to the interaction of many components and variability of parameter values causing radically different dynamics. The modelling process can be enhanced through the use of data mining techniques. We demonstrate this by applying association rules and clustering techniques to two stages of modelling: identifying pertinent structures in the initial model creation stage, and choosing optimal parameters to match that model to observed data. This is illustrated through application to the study of the circulating mumps virus in Scotland, 2004-2015.
Friday
29 Sept
Dr. Sandy Brownlee Planes, training and optimobiles: adding value to optimisation in the real world   Optimisation problems can be found nearly anywhere. There is always something that can be made greener, faster or more efficient. With a little machine learning, a lot of metaheuristics, some exact methods and a bit of visualisation, we can provide near-optimal solutions to real-world problems, but better still we can explain how we got there, and why the solutions are the right ones. This talk will focus on some ongoing work I have in modelling and optimisation of taxiing aircraft, and in optimal building design. I'll also touch on a few other application areas I'm interested in - all of which are about "adding value" to optimisation.
Friday
3rd Nov
Dr. Paul McMenemy The Sphere-Packing Bound and Travelling Salesmen - What Do They Have in Common?   The answer is `very little', except for the fact that they are both topics that will be covered in this seminar. The Travelling Salesman Problem (TSP) is one of the best known combinatorial optimisation problems, with the Chained Lin-Kernighan (CLK) heuristic being popularly used to obtain approximate solutions for TSP since it was proposed in the 1970s. CLK is an iterated local search heuristic that employs a double-bridge kick to escape from a stalled local search, adaptively relocating to a new location to recommence a local search for improved fitness values of the TSP problem. I present here an analysis of how the strength of the double-bridge kick used by CLK impacts upon the effectiveness of the CLK heuristic in finding the a priori global optimum of TSP instances, specifically analysing random and clustered TSP instances of increasing sizes. Binary codes are utilised by all modern information and computational systems, where they are used to transmit information using many different protocols. Some binary codes provide alphabets of codewords that inherently contain error-detecting and correcting capabilities. However, the number of codewords in these alphabets (M) are limited by a combination of the length of the binary string used (n), and the Hamming distance (d) used to enforce the levels of error detection and correction required. The maximum number of codewords that these binary (n,M,d)-codes can utilise are upper-limited by the Sphere-Packing (or Hamming) Bound value. I propose a method for determining the maximum number of codewords that a binary (n,M,d)-code can produce, and show some results that tentatively indicate a relationship between the upper limit of the Sphere-Packing Bound and the maximum number of codewords generated by this novel method.
Friday
10th Nov
Prof. Quintin Cutts Room 2B88 Does Lego help develop CS skills?   This talk describes a study to examine possible links between early childhood activities and later success with CS and IT. As we explore appropriate curricula for an academic discipline of CS, as we surely must do when introducing mandatory CS education into primary schools, such links are important. Could they explain oft-reported high failure rates in our subject? Can they be used to create an effective progression framework for CS education, to augment current thinking on the topic? Will we identify early years and primary activities that could be of value in other subjects as well? One such is the development of spatial skills, shown by researchers to be associated with success in STEM subjects, but which appears to have received little attention in mainstream education.
Friday
17th Nov
Dr. Fiona McNeill Engaging children and working towards gender equality in Computer Science   This is a two-part talk. In the first part I will talk about my work in school outreach, and particularly my involvement in the First Lego League - a robotics competition for 9-16 year olds which we have been running across Scotland for that last 5 years. I will discuss what we have been doing to encourage engagement from girls and from children from disadvantaged areas, and to ensure long-term impact from the programme. In the second part I will talk about my work around gender equality, including a new working group I am involved in that is looking back at the Tapping all our Talents report (RSE, 2012) to see how we have progressed, and work I am doing with our female students at Heriot-Watt. Hopefully this will lead on to discussions about best practice in supporting female students, drawn from what is being done in Heriot-Watt and Stirling, as well as any other environment people have experience of.
Friday
1 Dec
Dr. Nada Veerapen Exploring Search-Based Software Engineering Fitness Landscapes  Broadly speaking, Search-Based Software Engineering is concerned with the application of metaheuristic search techniques to optimise software engineering problems, from software requirements engineering to testing and bug fixing. This talk considers Genetic Improvement in particular, which modifies existing source code to make it `better', usually by fixing bugs or by improving some non-functional property such as execution time or power consumption. Such problems are optimisation problems that generally cannot be solved exactly. A search algorithm has to explore what is commonly referred to as the problem?s fitness landscape to find a good solution. This metaphor becomes rather abstract when working with multidimensional spaces. The talk will look at Local Optima Networks as a model of the global structure of the search space and how dimensionality reduction can help visualise the corresponding landscapes.

Previous Seminar Series

2023:  Spring   Autumn
2022:  Spring   Autumn
2021:  Spring   Autumn
2020:  Spring   Autumn
2019:  Spring   Autumn
2018:  Spring   Autumn
2017:  Spring   Autumn
2016:  Spring   Autumn
2015:  Spring   Autumn
2014:  Spring   Autumn
2013:  Spring   Autumn
2012:  Spring   Autumn
2011:  Spring   Autumn
2010:  Spring   Autumn
2009:  Spring   Autumn
2008:  Spring   Autumn
2007:  Spring   Autumn
2006:  Spring   Autumn
2005:  Spring   Autumn
2004:  Spring   Autumn
2003:  Spring   Autumn
2002:  Spring   Autumn
2001:  Spring   Autumn
2000:  Spring   Autumn
1999:  Spring   Autumn
1998:  Spring   Autumn
1997:  Spring   Autumn
1996:  Autumn
 

Top image: Illustrated example of running the Epsilon-constraint algorithm in order to maximise two objectives: find an optimal solution for objective 1; restrict the solution space according to the solution's value for objective 2 and look for an optimum solution of objective 1 in that space; repeat the previous step until there are no more solutions to be found. Any dominated solutions need to be filtered out of the set of solutions.
Courtesy of Dr. Nadarajen Veerapen. Related to a recent publication:

N. Veerapen, G. Ochoa, M. Harman and E. K. Burke. An Integer Linear Programming approach to the single and bi-objective Next Release Problem. Information and Software Technology, Volume 65, September 2015, Pages 1-13, ISSN 0950-5849. DOI:10.1016/j.infsof.2015.03.008


This page is maintained by:
Computing Science and Mathematics
Faculty of Natural Sciences
Room 4B102, Cottrell Building
University of Stirling, Stirling FK9 4LA
Tel: +44 1786 46 7286


© University of Stirling FK9 4LA Scotland UK • Telephone +44 1786 473171 • Scottish Charity No SC011159
Portal Logon

Forgotten login?

×