Research

Awarding Excellence: Smart & Sustainable IT for IEEE 2025 World Forum on Internet of Things: Dr Di Wu

Chair of the IEEE IoT Educational Activities Committee and Award Presenter, Dr Wanqing Tu, alongside Di Wu

From St Andrews to Chengdu, Dr. Di Wu has been awarded third place in the 2025 IEEE World Forum on Internet of Things PhD Thesis Competition.  IEEE is an internationally recognised organisation within electrical and electronics engineering. By participating in such an event, participants can receive valuable external feedback and connect with a larger community focused on the future vision of IoT Systems.

With a thesis titled “Distributed Machine Learning on Edge Computing Systems,” Dr. Wu proposes three techniques to better train machine learning models that directly affect small devices such as sensors, smartphones, and every day IoT gadgets. He states that the focus on smaller devices is becoming even more important due to the grand size of modern datasets, as well as how time-consuming, expensive, and at-risk to user privacy sending information to the cloud can be:

In my research, I proposed three techniques to make this kind of training more practical. The first helps devices decide how to split and share the workload. The second reduces the amount of data that needs to be exchanged during training. And the third lowers the amount of computation each device has to perform. Finally, I brought all these ideas together into one complete system. When we tested it on real IoT devices, it trained models faster, communicated less data, and achieved better accuracy compared with existing methods.

This improvement in efficiency suitably aligns with IEEE’S 2025 theme of “Smart and Sustainable IoT.” ‘To me’ Dr Wu states, ‘“smart” IoT means giving devices the ability to learn and make decisions locally. While “sustainable” IoT means doing this in a way that saves energy, protects user privacy, and can scale as the number of devices continues to grow. Therefore, by cutting down the computation and communication needed for training, intelligent IoT systems can become more sustainable and easier to deploy in practice.’ With this in mind, Dr. Wu propelled forward with his research that was also greatly influenced by the challenges he experienced as a machine learning engineer and the specific research questions that arose from reading subject-specific literature, discussing ideas with his supervisor Blesson Varghese, as well as building real-world prototypes throughout his PhD journey.

I truly see preparing for the nomination as a natural step that came out of the work I did during my PhD. I had published papers in related venues, including the IEEE Internet of Things Journal and IEEE Transactions on Parallel and Distributed Systems, which gave me some confidence that my work was heading in the right direction. Furthermore, writing my thesis, presenting ideas at conferences, as well as preparing for my viva helped me clarify my ideas which eventually helped me piece together and highlight the parts of my research that were most relevant to the theme. I would really encourage PhD graduates to apply for these kinds of thesis competitions.[1]

Now working as a Research Fellow funded by the UK National Edge AI Hub, Dr. Wu reflects on how this year’s IEEE displayed active research engagement with the intersection of AI and IoT — ‘both AI for IoT, where AI is used to solve IoT-specific problems, and AI on IoT, where we try to bring AI capabilities directly onto IoT devices.’ Another emerging direction he noted was the integration of sensing, communication, and computation. ‘These used to be relatively separate research areas, each led by different communities. But now we’re seeing growing interest in combining them into a single, unified system, which I think has a lot of potential.’ As Dr. Wu continues to explore efficient and scalable machine learning systems at the edge, he believes his new research direction will move beyond traditional federated learning, turning specifically to how agent-based systems and efficient foundation models (such as large language models) can be brought to the edge. ‘These areas are quite different from conventional ML systems, but they open up exciting possibilities for the next generation of edge intelligence,’ he concludes.

[1] Dr. Di Wu personally recommends competitions such as, ACM PhD Competition, the IEEE IoT PhD Competition, the IEEE TCSC PhD Thesis Award, as well as local competitions like the SICSA PhD Competition in Scotland.

PGR Seminar: Ben Claydon and Joseph Loughney

You are warmly invited to the next PGR Seminar.

Tear off calendar Monday 17/11/2025

Timer clock 14:00-15:00

Pin JC 1.33A

  1. Speaker: Ben Claydon

Title: Improvements to Space Partitioning Trees for Similarity Search

Abstract: Searching large, unstructured collections of data for objects deemed of relevance to a user-provided query is an increasingly important task. For example, when a user inputs a query into a search engine, they expect a list of high-quality results to be returned nearly immediately. However, if the search engine had to rank each of the approximately 1 billion websites on the internet in order of relevance to your query, and each of these judgements took 1 microsecond, each web search would take nearly half an hour! To overcome this, data structures are created which enables searching only a small subset of the dataset, where this subset is likely to contain the most relevant items. This greatly increases the rate at which queries can be served, at the cost of some loss of accuracy. In this talk, I will present improvements made to one such data structure called random projection forests. I present a method which increases the accuracy of this algorithm without an associated increase in either preprocessing time or the time required to serve a query.

Bio: Ben is a 3rd year PhD student whose research focusses on algorithms which facilitate scalable similarity search.

  1. Speaker: Joseph Loughney

Title: Breaking Multiple Symmetries at Once, and Retrieving ‘Actual’ Solution Counts in Graph Search Problems

Abstract: The Subgraph Isomorphism Problem is a graph search problem which is computationally challenging to solve in large cases. We have seen significant improvements in performance as a result of symmetry breaking (avoiding searching for multiple solutions that are isomorphic to each other) either before or during search, and on either the pattern graph, the target graph, or both. Can we use any of these techniques in combination with each other? What about when counting solutions? How can we return the ‘actual’ number of solutions from the number of equivalence classes? This talk will aim to approach some answers to these questions.

We hope you can join us!

PGR Seminar: Charis Hanna and Maria Andrei

You are warmly invited to the next PGR Seminar.

Date & Time: Monday 10/11/2025 14:00-15:00

Location: JC 1.33A

  1. Speaker: Charis Hanna

Title: Self-Supervised Learning for Efficient Ecological Monitoring

Abstract: Cliff-nesting birds serve as valuable indicators of marine ecosystem health, yet dense populations and remote habitats present significant challenges for automated monitoring. With current state-of-the-art object detectors often failing under the conditions of extreme crowding and occlusion, this project aims to develop and refine deep learning techniques that enable the fine-grained, automated analysis of seabird colonies. Current work explores self-supervised learning strategies that leverage domain-shifted knowledge to reduce the need for exhaustive annotation across complex image datasets. These methods not only reduce the laborious process of manual annotation but also demonstrate promising improvements in performance across the long-tailed species distribution. While ongoing efforts are directed at further optimising these models, future work will leverage additional spatial information with the aim of supporting richer insights into behavioural dynamics within these populations.

Bio: Charis is a 3rd-year PhD student developing novel deep learning approaches for the automated monitoring of dense cliff-nesting bird colonies. Her research focuses on advancing computer vision methods for detection, classification, and behavioural analysis in challenging habitats.

  1. Speaker: Maria Andrei

Title: Bridging Psychological Distance from Climate Change through Experiential Learning within Heritage Organisations

Abstract: Climate change represents one of the most urgent challenges of our time, not only in its environmental impacts but also as a complex science communication problem. Despite broad scientific consensus on its causes and mitigation pathways, public understanding and engagement remain fragmented, limiting the collective action needed to address this crisis. My research investigates how immersive technologies, particularly virtual reality, can bridge the gap between scientific knowledge and public perception by transforming abstract climate data into tangible, emotionally resonant experiences. By connecting global and local climate futures through case studies such as Antarctica and Scotland, I examine how immersive simulations can reduce psychological distance from climate change. By evaluating audience responses across diverse contexts, from museums to polar expedition vessels, this research aims to assess how experiential storytelling can improve climate communication and motivate engagement with climate action.

Bio: Maria is a third-year PhD researcher working with the Schools of Computer Science, Biology, and Earth & Environmental Sciences. Her work focuses on immersive climate communication, using virtual reality to visualise climate impacts in regions such as Scotland and Antarctica. She collaborates with heritage organisations, science centres, and polar expedition companies to bring these experiences to communities across Scotland and beyond.

We hope you can join us!

PGR Seminar – Erdem Kus & Junyu Zhang

You are warmly invited to the next PGR Seminar.

Date & Time: Monday 20/10/2025 14:00-15:00

Location: JC 1.33A

  1. Speaker: Erdem Kus

Title: Frugal Algorithm Selection for Combinatorial Search

Abstract: Solvers for combinatorial search and optimisation problems often exhibit highly complementary performance: instances that are hard for one solver may be easy for another. The Algorithm Selection Problem (ASP) addresses this by predicting, for each problem instance, which solver will perform best. Machine learning models trained for this purpose, however, are typically expensive to construct, as they require exhaustive solver runs on all training instances to obtain ground-truth performance data.

In this work, we propose a frugal alternative that formulates algorithm selection as an active learning problem. Instead of uniformly evaluating all solver–instance pairs, our method intelligently selects the most informative ones, thereby drastically reducing the cost of data collection. We show that standard active learning techniques are inadequate for this setting, as they overlook the structure and cost characteristics unique to algorithm selection. To address this, we introduce novel, cost-aware active learning strategies that leverage auxiliary models to balance informativeness and evaluation cost.

Bio: Erdem is a PhD candidate whose research focuses on Artificial Intelligence (AI) and Constraint Programming (CP).

  1. Speaker: Junyu Zhang

Title: Remaking Characters in Heritage Contexts to Support Inclusive Learning

Abstract: Characters in immersive environments have the potential to enrich user experience, improving engagement with heritage and in so doing benefiting heritage organisations and their communities. Creating authentic digital scenes based upon survey, archaeological and historical data, co-creative design and community engagement enables communities and their visitors to understand the past better. The understanding of authenticity stimulates the potential of enriching cultural heritage with the details of lives past and also discusses how this research benefits the Sustainable Development Goals.

Bio: Minty is a PhD candidate exploring the authenticity of characters to support inclusive learning in heritage contexts. She is interested in how digital technologies can be used in the intersection of different disciplines to achieve SDGs in the field of cultural heritage, so as to enhance the promotion, representation, and well-being in digital humanities education and also affect resonated dialogue and thinking among diverse people and communities in facing the current challenges.

We hope you can join us!

PGR Seminar – Qurat ul ain Shaheen

You are warmly invited to the next PRG Seminar.

Date & Time: Monday 13/10/2025 14:00-14:40

Location: JC 1.33A

Speaker: Qurat ul ain Shaheen

Title: A Framework for Uncertainty Sampling in Active Learning

Abstract: Uncertainty sampling is an active learning paradigm where data instances representing maximum uncertainty for a machine learning model are selected for training. This talk will explore existing uncertainty modelling approaches for binary classification of categorical data.  It will introduce a conceptual framework to improve uncertainty modelling and present some preliminary results.

Bio: Qurat ul ain Shaheen is a final year PhD researcher. Her research focuses on modelling uncertainty in active learning.

We hope you can join us!

Young Software Engineer of the Year 2025 Awards

Huge congratulations to Verity Powel, a winner at last night’s Young Software Engineer of the Year Awards (https://www.scotlandis.com/blog/rugby-video-tech-scores-top-award-for-st-andrews-student/). Her final year project “Video Analytics For Rugby Skills Training” was nominated by the school (https://blogs.cs.st-andrews.ac.uk/csblog/2025/07/28/nomination-to-young-software-engineering-of-the-year-awards-2025/) in June. The awards were announced at the ScotSoft 2025 (https://www.scotlandis.com/scotsoft-2025/), Scotland’s leading tech conference at the Edinburgh International Conference Centre.

The Young Software Engineer of the Year accolades are awarded to the best undergraduate software projects from students studying computer science and software engineering in Scotland. Over the years, St Andrews has many finalists and prize winners.

PGR Seminar – David Morrison

You are warmly invited to the next PRG Seminar.

Date & Time: Monday 06/10/2025 14:00-14:40

Location: JC 1.33A

Speaker: David Morrison

Title: Synthetic Whole Slide Image Patch Embeddings for Multiple Instance Learning

Abstract: Obtaining high-quality data is a persistent challenge for the training of computational pathology models. As medical data, Whole-slide images (WSIs) are often held under restrictive terms by medical institutions and, as a result, are hard to access by researchers. Where data is available, the number of whole slide images can be limited and skewed towards common pathology types. In addition, there can be issues with labelling: slide-level labels may lack information about specific pathologies, for example, they may be limited to binary labels of normal or malignant, while annotations at the level of patches are rarely available.

Synthetic data generation is a possible solution to these problems by allowing researchers to produce data on demand that can be used in an unrestricted manner with high-quality labels. I have previously presented on the generation of synthetic patch data. In this talk, I will discuss an extension to this work in which this approach is combined with models trained to characterise the slide as a whole in order to provide a synthesis process for data for use with multiple instance learning techniques, commonly used in whole slide image classification.

We hope you can join us!

 

PGR Seminar – Sharon Pisani & Mirza Hossain

The next PGR seminar is taking place this Friday 3rd October 11:00-12:00 in JC 1.33A.

Below are the Titles and Abstracts for Sharon and Mirza’s talks – Please do come along if you are able.

Sharon Pisani

Title: Building Sustainable Heritage Virtual Museums for Communities using Sociodata

Abstract: Virtual museums are moving beyond simple digitisation of artefacts to become dynamic platforms for community engagement and sustainable development. This talk introduces the VERA Platform, which combines a flexible Virtual Museum Infrastructure with a new layer of sustainability-oriented contextual data called sociodata. Sociodata links heritage objects to their cultural landscapes, local communities, and relevant Sustainable Development Goals, enabling richer discovery, analysis, and reuse. In this talk, I will outline the platform’s architecture and metadata model. The talk will highlight technical challenges such as interoperability with European data spaces, and supporting interactive storytelling at scale—issues highly relevant to digital infrastructure and data-driven research in the heritage sector.

Bio: Sharon is a PhD researcher examining the role of emergent digital technologies in preserving and engaging with cultural heritage while supporting sustainable development. Her research focuses on digitising cultural landscapes—both natural and cultural heritage—to assess various impacts on heritage and community identities. She explores how digital tools, including 3D scanning, 3D modeling, and mixed reality, can aid in recreating and safeguarding heritage at risk.

Mirza Hossain

Title: Fishing for monosemantic neurons in histopathology foundation models

Abstract: This early-stage study introduces Histoscope, an interactive system for examining sparse autoencoders (SAEs) that are trained on top of the UNI pathology encoder. Vision transformers for histopathology often exhibit superposition, where single neurons respond to multiple distinct tissue patterns, making interpretation difficult. Histoscope provides quantitative metrics and visualisations to assess whether neurons are monosemantic—associated with a single concept—or polysemantic—associated with multiple concepts. The work highlights methods for analysing internal representations of histopathology foundation models and contributes to efforts toward more transparent AI in pathology.

Bio: Mirza Hossain is a second-year PhD candidate in Computer Science at the University of St Andrews. His research focuses on multimodal AI in medical imaging with an emphasis on mechanistic interpretability of large foundation models. He is supervised by Dr. David Harris-Birtill.