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 here.
| Date | Speaker | Title/Abstract |
|---|---|---|
| Friday 18th January |
Lucia Vadicamo, Graduate Fellow, Institute of Information Science and Technologies, Pisa | Improving Metric Search through Finite Isometric Embeddings Metric search is concerned with the efficient evaluation of proximity queries in metric spaces. Most metric indexing and searching mechanisms rely upon the triangle inequality property. This property allows deriving bounds on the distance between data objects, which are used to efficiently exclude partitions of the data that do not contain solutions to a given query. This seminar will discuss a class of metric spaces meeting a stronger property, named n-point property, which is a condition defined in term of finite isometric embeddings into the Euclidean space. This property gives stronger geometric guarantees, in particular one named the Hilbert Exclusion property. Moreover, it can be used to embed metric objects into a finite-dimensional Euclidean space, which turns out to have implications not only for the similarity search but also for dimensionality reduction and data visualization tasks. Bio: Lucia Vadicamo graduated in Mathematics at the University of Pisa, Italy, in 2013 and was awarded a PhD in Information Engineering at the University of Pisa, Italy, in 2018. Her current research interests involve content-based image retrieval, similarity search and permutation-based indexing. |
| Friday 25th January |
Prof. Rachel Norman, CSM, University of Stirling | The Professional Doctorate (non-research talk) This session is aimed at those who are potentially interested in doing the Professional Doctorate in Big Data Science. This is a more applied form of a PhD and can follow on from the MScs run in the division. This talk is mainly aimed at MSc students and undergraduates who are interested in this possibility. It will be an informal session with the opportunity to ask questions. |
| Friday 1st February |
Prof. Paul Lambert, Sociology, Social Policy & Criminology, University of Stirling | Social science and social stratification: Research themes and challenges In this session I will discuss selected themes that currently occupy a prominent position in one important area of social science research. A brief overview of the 'Social Surveys and Social Statistics' research group in the Faculty of Social Sciences will first be given. Much, though not all, of the work in that group addresses the analysis of structured empirical data about enduring systems of social inequality (or 'social stratification'). It will be argued that four methodological issues in that domain are especially important in shaping current research endeavours and opportunities. Two are about exploiting data - how social scientists confront complex categorical data; and how they make use of data on social connections. Two more are traditionally associated with analytical methods - how we approach the analysis of causal relationships; and how we appropriately convey uncertainty. Taken together, these issues influence current research agendas within 'quantitative social science' on social stratification, but they also have interesting potential for outreach and impact beyond that academic niche. |
| Wednesday 13th February Note different day! Still 3pm, 4B96 |
Jussi Korpela, University of Helsinki, Finland | Focusing waves in unknown media in energy norm We study the wave equation on a bounded domain R^m or on a compact Riemannian manifold with boundary. Let us assume that we do not know the coefficients of the wave equation but are only given the hyperbolic Neumann-to-Dirichlet map that corresponds to physical measurements on a part of the boundary. We show that it is possible to construct a sequence of Neumann boundary values so that at a time t_0 the corresponding waves converge to zero while the time derivative of the waves converge to a delta distribution. A key feature of this result is that it does not require knowledge of the coefficients in the wave equation, that is, of the material parameters inside the media. However, we assume that the point where the energy of wave focuses is known in travel time coordinates, and satisfies a certain geometrical condition. |
| Friday 22nd February |
Mid Semester break | No seminar |
| Wednesday 27th February Note: different day and venue Room 2B48 |
Daniel Otero, Department of Mathematics, Xavier University, USA | Transforming instruction in undergraduate Mathematics via primary historical sources The speaker is one of a team of seven mathematicians and mathematics educators, representing different universities across the United States, who have been at work to design, author, classroom test, revise, evaluate, and disseminate classroom modules called Primary Source Projects (PSPs), which are meant to teach standard topics from across the early years of the undergraduate mathematics curriculum through primary historical source materials. This endeavor, called by the acronym TRIUMPHS, intends for PSPs to replace traditional textbook presentations of mathematical content by focusing student attention on the interpretation of historical source texts combined with brief contextual material and carefully crafted exercises meant to encourage sense-making by students. PSPs are also designed to incorporate principles of active learning, wherein the bulk of classroom time is given over to student work on project tasks and exercises, both alone and in discussion with small groups of classmates, or involving the entire class, rather than to traditional lectures by the instructor. The TRIUMPHS team, supported with funding from the US National Science Foundation, have created some 48 such modules together with a few external authors. These are now freely available from the TRIUMPHS website. Some PSPs can take as little as 30 minutes to implement, while others are designed to take up to four weeks (with median implementation time of about one week) of classroom time. There are modules written to support standard coursework from precalculus and calculus, linear algebra, differential equations, algebra, theory of numbers, geometry, analysis, statistics, and a few other subjects as well. This talk will discuss the TRIUMPHS endeavor generally but will show examples of PSPs at work through a pair of projects authored by the speaker, one of which is an introduction to the study of trigonometry, the other of which teaches the matrix determinant. Speaker Bio: Daniel E. Otero is Associate Professor of mathematics at Xavier University in Cincinnati, OH. His chief interests are in the history of mathematics and its uses in teaching, especially via primary sources. He is the most recent former Chair of HOM SIGMAA (the History of Mathematics Special Interest Group of the Mathematical Association of America) and was President of the Ohio Section of the MAA (2015-2016). He is currently a Visiting Fellow in the History of Mathematics at the University of St. Andrews. |
| Friday 1st March | Saemundur Haraldsson, Lancaster University | FIXIE Exploiting Defect Prediction for Automatic Software Repair The FIXIE project is a three year project funded by the EPSRC and it has only been ongoing for a few months. The goal of the project is to establish a new technique to automatically fix predicted defects in software code before testing. We will use machine learning-based defect prediction information to generate automatic fixes using Genetic Improvement. The approach aims to offer developers effective fixes to code which is predicted as defective. A higher proportion of the fixes our approach offers to developers should be acceptable, generated quicker and available earlier in the development cycle than previous attempts at automated repair. Importantly, our approach targets a wider pool of defects as it specifically includes targeting those dormant defects which are not identified by testing. |
| Friday 8th March Note: different venue Room 2B129 |
Patrick Maier, Computer Science and Software Engineering, Sheffield Hallam University | Scaling Parallel Combinatorial Search Combinatorial search is central to many AI applications such as planning, scheduling, natural language processing, automated reasoning and computational algebra. Such applications would benefit from search taking advantage of increasingly available large-scale parallel hardware resources such as cloud and HPC services. By its nature, combinatorial search has huge potential for parallelism, yet it is very difficult to scale due to extreme irregularity and algorithmic features (such as search order heuristics and learning) that interfere with parallelism. In this talk I will discuss the challenges of scaling parallel combinatorial search, particularly irregularity and search order heuristics. I will present a novel framework for parallel combinatorial search that addresses these challenges. The framework provides a skeleton-based programming model that offers users a low-cost route to parallelising combinatorial search applications. Results demonstrate that the framework can deliver good parallel performance, scaling state-of-the-art combinatorial search algorithms to hundreds of cores, achieving good speedups and low variance of parallel runtimes. BIO: Patrick Maier is a Lecturer in Computer Science and Software Engineering at Sheffield Hallam University. Prior to joining Sheffield Hallam in 2018, he held research positions at the Universities of Edinburgh and Glasgow, at Heriot-Watt University, and at the Max-Planck Institute for Computer Science in Saarbruecken, Germany, where he obtained his PhD in 2003. Patrick's research is concerned with programming languages and systems for parallel computation, specifically for large-scale problems with irregular parallelism. |
| Wednesday 13th March Note different day! Still 3pm, 4B96 |
David Li, Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde | Computational challenges and opportunities for advanced imaging techniques
Rapid advances in photonics and microelectronics have promised scientists to image biological processes with special resolution as tiny as a few nanometers or to capture ultrafast phenomena happening within nanoseconds. However, such advanced imaging systems also bring unprecedented challenges in data storage and data analytics. How can we deal with such challenges with innovative computational techniques and what opportunities will be promised through these new techniques? Speaker Bio: David Li is a senior lecturer in photonics/electronics. He received his PhD in electrical engineering from the National Taiwan University. He then joined the Industrial Technology Research Institute (ITRI), Taiwan, as a R&D engineer working on 1.25-12.5Gbps optical communication chipsets and wireless communication IP, knowledge transfer, and international joint projects with the Carnegie Mellon University, Pittsburgh, USA. From 2007 to 2011, he worked at the Institute for Integrated Micro and Nano Systems, University of Edinburgh, on EU "MEGAFRAME" and "METOXIA" projects for CMOS single-photon sensor cameras and analogue front-end circuits. From January 2014, he joined the Centre for Biophotonics, Strathclyde Institute of Pharmacy & Biological Sciences, after his first lecturing position in biomedical engineering and embedded systems at the School of Engineering and Informatics, University of Sussex, where he led a research team working on industry-funded projects. His current research interests include CMOS imaging and sensor systems, embedded systems and GPU computing, digital signal processing, mixed-signal integrated circuits, fluorescence based sensing systems, electrical impedance sensing systems, forward models of electrical impedance tomography and numerical modelling. He has been working with researchers within the UK and from oversea such as Netherlands, Switzerland, Belgium, Germany, France, China, USA, and Taiwan. |
| Friday 15th March |
Armando Marino, BES, University of Stirling | Polarised eyes in the sky: processing of polarimetric radar data It is undeniable that the world we live in is constantly monitored by a fleet of electronic eyes in the sky. However, the objectives of such instruments are rather different from the beliefs supported in some blogs. In the last decades, a large variety of sensors were built, able to remotely measure bio-physical parameters of our Planet as well as other Planets exploiting orbiting satellites. In this talk, I would like to take you for a stroll in my attempts of using mathematical tools and signal processing “tricks” to extract information from polarimetric radar data. I will show you how we can design statistical tests and apply optimisations or solve integrals of quadratic forms to manipulate the image pixels for a large variety of applications, from detecting icebergs to measuring the height of trees in forests. Speaker Bio: Armando Marino received his MSc in Telecommunication Engineering at the Universita’ di Napoli “Federico II” in 2006. In 2006, he joined the Microwaves and Radar Institute, German Aerospace Centre (DLR), Oberpfaffenhofen, where he developed his MSc thesis, which focused on SAR multi-pass retrieval of forest parameters. In 2011, he was awarded his PhD degree at the University of Edinburgh, (School of Geosciences), Edinburgh, UK in the field of polarimetric SAR interferometry. His PhD thesis was awarded “Best PhD Thesis 2011” by the RSPSoc (Remote Sensing and Photogrammetry Society) and “Outstanding PhD Thesis” by Springer Verlag, which published the thesis in 2012. From March 2011 to October 2011, he was working within the University of Alicante, Institute of Computing Research, Spain. From December 2011 to May 2015, he was a post-doctorate researcher and lecturer at the ETH Zürich, Institute of Environmental Engineering, Switzerland. In June 2015 he joined as a lecturer the Open University, Engineering and Innovation, Milton Keynes, United Kingdom. Since April 2018 he is a senior lecturer at the University of Stirling, Biological and Environmental Sciences. His main research interests include processing of stacks of radar images for surveillance and parameter retrieval, developing ground based radars and applying Computer Vision and Machine Learning methodologies to remote sensing images. |
| Friday 22nd March |
David White and Benjamin Jones, University of Sheffield | Software Optimisation for Quantum Computing Quantum Computing hardware is developing rapidly, and we anticipate useful machines capable of outperforming classical computers for certain tasks within the next few years. Due to the difficulties of constructing, manipulating, and preserving quantum bits (qubits), the non-functional properties of software running on quantum computers is of great importance. I will give a brief introduction to quantum computing, explain why non-functional software properties matter, and describe an example: the optimisation of physics simulations for quantum computers. I hope that this talk will encourage other researchers to apply their expertise to quantum software! There is much work to be done. This talk also includes work by John A. Clark, George O'Brien and Earl Campbell. All Sheffield; Earl in Physics, John and George in CS. |
| Friday 29th March |
Yulia Timofeeva, University of Warwick | Action potential counting at giant mossy fiber presynaptic terminals:
experiments and computational modelling Neurons fire action potentials to transfer information through synaptic release of neurotransmitter. At presynaptic terminals, the pattern of action potential discharge is integrated through dynamic Ca2+ signalling by the presynaptic machinery which triggers the release of neurotransmitter. It is generally accepted that the rate and the temporal precision of action potential firing support information transfer between neurons. Here, we show that in contrast to rate and temporal coding, giant mossy fiber terminals count the number of action potentials during trains to trigger CA3 pyramidal cell firing. Our results shed light on the synaptic signal transfer mechanisms supporting an additional information coding strategy in the brain. In our study we combined electrophysiological recordings with rapid presynaptic two-photon Ca2+ imaging and experimentally constrained modelling. Speaker Bio: Yulia Timofeeva is an Associate Professor in the Department of Computer Science at the University of Warwick. She is also linked at Warwick with the Centre for Complexity Science, the Systems Biology & Infectious Disease Epidemiology Research (SBIDER) group and the Mathematics for Real-World Systems Centre for Doctoral Training. Prior to joining Warwick in 2007, she held research positions at the University of Nottingham and Heriot-Watt University. The main area of her research is focussed on understanding cellular signalling mechanisms using mathematical and computational modelling. Her close collaboration with neurophysiologists at UCL led to her recent additional affiliation as an Honorary Senior Research Associate in the Department of Clinical & Experimental Epilepsy, UCL Queen Square Institute of Neurology, London. Yulia's background includes an undergraduate degree in Applied Mathematics (St.Petersburg, Russia), MSc in Industrial Mathematical Modelling (Loughborough University) and PhD in Mathematical Biology (Loughborough University). |
| Friday 5th April |
Internal events | |
| Friday 12th April |
Prof Lei Tang, Department of Electrical Engineering, Hefei Normal University, China | High temperature detection and control technology The detection of temperature and combustion efficiency in ethylene cracking furnaces has always been a key problem to be solved in the petrochemical industry. We will introduce the detection methods of high temperature and combustion efficiency that seeks advanced control methods based on accurate detection to solve the current over-oxygen combustion control problems for ethylene cracking furnace. |
| Friday 19th April |
Good Friday holiday - university closed | |
| Friday 26th April |
Available slot | |
| Friday 3rd May |
Ian Gent, School of Computer Science, University of St Andrews | The winnability of Klondike and many other single-player card games The most famous single-player card game is ‘Klondike’, but our ignorance of its winnability percentage has been called “one of the embarrassments of applied mathematics”. Klondike is just one of many single-player card games, generically called ‘solitaire’ or ‘patience’ games, for which players have long wanted to know how likely a particular game is to be winnable for a random deal. A number of different games have been studied empirically in the academic literature and by non-academic enthusiasts. Here we show that a single general purpose Artificial Intelligence program, called “Solvitaire”, can be used to determine the winnability percentage of approximately 30 different single-player card games with a 95% confidence interval of ± 0.1% or better. For example, we report the winnability of Klondike to within 0.10% (in the ‘thoughtful’ variant where the player knows the location of all cards). This is a 30-fold reduction in confidence interval, and almost all our results are either entirely new or represent significant improvements on previous knowledge. Speaker Bio: Ian Gent is professor of Computer Science at the University of St Andrews. His mother taught him to play patience and herself showed endless patience when he “helped” her by taking complete control of the game. A program to play a patience game was one of the programs he wrote on his 1982 Sinclair Spectrum now on the wall outside his office. |
| Friday 10th May |
Stewart Blakeway, School of Computer Science, University of Manchester | An investigation of Mobile Ad hoc Network Performance with Cognitive Attributes Applied Mobile Ad hoc Networks (MANETs) are known for their versatility, they are capable of supporting many applications and are quick to deploy without need for an existing predefined communications infrastructure. However, although the lack of infrastructure allows for the quick deployment of the data communications network, it adds many factors that hinder packet delivery. Such hindrances occur because of the dynamic topology caused by the mobility of the nodes which results in link breakages. Routing protocols exist that attempt to refresh available routes; however, this is after link breakages have occurred. The nodes also usually have constrained resources (i.e. energy source and limited bandwidth). In our research we have adopted a novel approach of investigating network behaviour and management by implementing cognitive attributes into a MANET environment. This allows an application to better meet its mission objectives, decreases the end-to-end delay, and increases packet delivery ratio. The network can make observations, consider previous actions and consequences of the actions, and make changes based on the prior knowledge and experience. This work also shows how the network can better utilise limited resources such as bandwidth and drones by applying cognitive attributes. Simulations conducted show promising results and prove that an increase in network performance is possible if adopting a cross-layered approach and allow the network to manage and to ‘think’ for itself. Speaker Bio: Dr Stewart Blakeway was awarded his first two degrees by The University of Liverpool. He completed his PhD in Computer Science in 2015 at Liverpool John Moore’s University, investigating Adaptive Self-Learning Mobile Ad hoc Networks with cognition incorporated. Stewart’s research interests are in the optimization and analysis of mobile networks, the optimization of resource allocation and energy efficient networks. Stewart is currently collaborating with researchers in the UK, Russia and France. Stewart has over 20 years’ experience in Higher Education and is a Fellow of the Higher Education Academy (FHEA). Currently, his teaching is focused on Distributed Databases and Programming. |
| Friday 7th June |
Andrew Morozov, University of Leicester, United Kingdom | Towards constructing a mathematically rigorous framework for modelling evolutionary fitness In modelling biological evolution, a major mathematical challenge consists in an adequate quantification of selective advantages of species. Current approaches to modelling natural section are often based on the idea of maximization of a certain prescribed criterion - evolutionary fitness. This paradigm was inspired by the seminal Darwin's idea of the 'survival of the fittest'. However, the concept of evolutionary fitness is still somewhat vague, intuitive and is often subjective. On the other hand, by using different definitions of fitness one can predict conflicting evolutionary outcomes, which is obviously unfortunate. In this talk, I present a novel axiomatic approach to model natural selection in dynamical systems with inheritance in an arbitrary function space. For a generic self-replication system, I introduce a ranking order of inherited units following the underlying measure density dynamics. Using such ranking, it becomes possible to derive a generalized fitness function which maximization will predict long-term evolutionary outcome. The approach justifies the variational principle of determining evolutionarily stable behavioural strategies. I demonstrate a new technique allowing to derive evolutionary fitness for population models with structuring (e.g. in models with time delay) which was so far a mathematical challenge. Finally, I show how the method can be applied to a von Foerster continuous stage population model. Speaker Bio: Andrew Yu Morozov is currently Reader at the Department of Mathematics of the University of Leicester. Having obtained his PhD at the Russian Academy of Sciences he has held research positions at the Shirshov Institute of Oceonology, University of California Riverside, as well as at the INRA Bordeaux. Andrew is Editorial Board member of several journals, and organised a number of prestigeous international meetings in mathematical biology. |
Top image: Glazed building facades along the trade-off of operational energy consumption for heating, lighting and cooling, vs capital construction cost, with a heatmap showing the result of mining a fitness model to identify ideal glazing locations.
Courtesy of Dr. Alexander Brownlee. Related to a recent publication:
Brownlee, A. E. I. Mining Markov Network Surrogates for Value Added Optimisation. Surrogate Assisted Evolutionary Optimisation (SAEOpt) Workshop in: Companion Proc. of the Genetic and Evolutionary Computation COnference 2016, Denver, CO, USA. DOI:10.1145/2908961.2931711