Computing Science and Mathematics Seminars,
Spring 2024
Unless otherwise state, seminars will take place in Room 4B96, Cottrell Building, University of Stirling from 13.00 to 14.00 on Friday afternoons, followed by informal discussions.
Spring 2024
| Date | Speaker | Title/Abstract |
|---|---|---|
|
Friday 2 February |
No seminar |
|
| Friday 9 February |
Paulius Stankaitis, CSM, University of Stirling | Synthetic Data for Artificial Intelligence Synthetic data has great potential in addressing data scarcity issues in the AI domain. In this talk, I will present ongoing work and discuss future directions on using realistic synthetic data for improving the trustworthiness of AI models (i.e., convolutional neural networks). The talk will describe a generic architecture which tries to generate additional synthetic images that lead to improved model accuracy and robustness. The talk will also describe some of the challenges, namely, the large state-space and image realism.
Speaker bio: Paulius Stankaitis is a Lecturer in AI/Data Science at the University of Stirling. He obtained his PhD degree from Newcastle University (United Kingdom) in the area of formal methods. His current research focuses on developing formal methods tools and techniques for trustworthy AI, cyber-physical systems and Digital Twins. He previously worked with companies such as Siemens Rail Automation, The Formal Route and Systra Scott Lister in developing and deploying practical formal methods to industry. |
| Friday 16 February |
Robin Hankin, CSM, University of Stirling | Friendly ghosts, Draw Monsters, and Lewis Hamilton: sports
statistics with the Bradley-Terry model In this short non-technical talk I discuss a method for analysing competitive situations including a range of sports and e-sports. I will demonstrate newly written software and analyse datasets drawn from a wide range of competitive situations such as:
Speaker bio: Robin Hankin is a mathematician specializing in computational statistics. His PhD is from the University of Cambridge in fluid mechanics, studying uncertainty in atmospheric diffusion. His first academic post was as a lecturer in environmental science, publishing in the field of industrial risk assessment. After returning to Cambridge to study uncertainty in global climate change as a senior research associate, he switched fields to computational statistics at Auckland University of Technology. His current research portfolio at the University of Stirling focusses on the analysis of sports statistics. |
| Friday 23 February |
Ana Lucia Garcia Pulido, CSM, University of Stirling | Machine meets theory: Lie algebras. This talk will begin with an informal introduction to Lie algebras and an overview of the field. We will then move on to one problem of interest, the classification of sympathetic Lie algebras. We will construct, step by step, a faithful combinatorial representation of this problem. This representation enables us to develop new theory and design new algorithms that, when combined, allowed us to completely solve the problem. This work is joint with G. Salgado. This talk will be accessible with a basic knowledge of linear algebra.
Speaker bio: Ana Lucia recently joined the Computing Science and Mathematics division as a Lecturer in Mathematics. Prior to this, she held postdoctoral positions at the University of Liverpool and at CIMAT (Guanajuato). She was awarded a PhD in Mathematics from the University of Warwick. |
| Friday 1 March |
No seminar |
|
| Friday 8 March |
No seminar (reading week) | |
| Friday 15 March |
Yijun Yan, Computing, University of Dundee | What Hyperspectral Imaging Can Do? Hyperspectral imaging (HSI) is an advanced imaging technique that captures and processes data from across the electromagnetic spectrum, going beyond the visible range to encompass ultraviolet and infrared light. Unlike conventional imaging that relies on three primary color channels—red, green, and blue—HSI incorporates potentially hundreds of spectral bands, offering intricate spectral information for each pixel. This comprehensive spectral data facilitates the precise identification and analysis of a wide array of materials and objects, distinguishing them based on their unique spectral signatures. This presentation delves into the fundamentals of HSI and explores its diverse applications across various fields such as remote sensing and non-destructive inspection, etc. Additionally, it highlights the potential and challenges associated with integrating AI methodologies in HSI analysis, offering insights into how these techniques can enhance data interpretation and application outcomes.
Speaker bio: Dr. Yijun Yan received his PhD in Electronic and Electrical Engineering University of Strathclyde (UoS), Glasgow, UK in 2018. After PhD, he worked as a postdoc in UoS and Robert Gordon University. He currently holds the position of Lecturer in Computing at the School of Science and Engineering, University of Dundee, alongside an honorary role as Lecturer at Robert Gordon University. His research primarily focuses on pattern recognition, hyperspectral imagery, and their application in remote sensing, as well as non-destructive testing across diverse industrial environments. Dr. Yan has contributed extensively to renowned journals such as Pattern Recognition, IEEE Transactions on Instrumentation and Measurement, IEEE Transactions on Geoscience and Remote Sensing, and IEEE Journal of Biomedical and Health Informatics. Several papers have been recognized as highly cited papers by the Essential Science Indicators (ESI). Furthermore, he serves as a Guest Editor for Remote Sensing and Frontiers in Plant Science and is a respected reviewer for prestigious journals including IEEE TIP, TGRS, and TIM, etc. Dr. Yan has been involved in numerous significant research projects funded by various organizations, including the European Union (EU-H2020), the UK Natural Environment Research Council (NERC), Innovate UK, the UK Office of Naval Research (ONR), and other international funding agencies and companies, with total funding exceeding £7 million. |
| Friday 22 March |
No seminar | |
| Friday 29 March |
No seminar (holiday) | |
| No seminar in April 2024 | ||
| Friday 3 May |
Irem Yaman, KBC Group | Credit Risk Modelling Credit risk models are used to estimate the losses that financial institutions would face when borrowers fail to meet their financial obligations. In this talk, we will discuss the intricacies of credit risk modelling and the regular assessment of using these models within IRB (Internal Ratings-Based) advanced banks—financial institutions authorised to use advanced models to assess credit risk under Basel II regulatory framework.
Speaker bio: Dr Irem Yaman, trained as a mathematician, has been actively involved in the modelling and validating credit risk models within the KBC Group, one of the largest bank-insurance groups in Belgium, as well as in Central and Eastern Europe. After obtaining her PhD from Istanbul Technical University in Turkey, where her research focused on rational approximation techniques, Dr Yaman pursued postdoctoral research at the University of Antwerp in Belgium, specialising in numerical computations of multivariate integrals. She was an assistant professor in the Mathematics department at Gebze Technical University in Turkey for three years, before moving back to Belgium and joining KBC in 2015. At KBC, Dr Yaman has held various roles, including the one in treasury, where she initiated the use of numerical techniques in identifying operational deposits in the books of KBC. In 2020, she set up a team of data engineers to provide automated solutions to credit risk validation activities within the Group. Currently, Dr Yaman is the department head of validation in KBC Global Services, Bulgaria branch. |
Previous Seminar Series
Computing Science and Mathematics
Faculty of Natural Sciences
University of Stirling
Stirling
FK9 4LA
Scotland
UK