Seminar

PGR Seminar with Dhananjay Saikumar

The next PGR seminar is taking place this Friday 25th April at 2PM in JC 1.33a

Below are the Title and Abstract for Dhananjay’s talk – Please do come along if you are able.

Title: Signal Collapse in One-Shot Pruning: When Sparse Models Fail to Distinguish Neural Representations

Abstract: The deep learning breakthrough in 2012, marked by AlexNet’s success on the ImageNet challenge, ushered in an era of increasingly large neural networks. Modern models now hold tens of millions to billions of parameters, enabling remarkable capabilities but creating serious challenges for deployment in real-world, resource-constrained environments. This has led to growing interest in model compression, with network pruning emerging as a widely adopted method to reduce computational and memory demands. Iterative pruning—based on repeated prune-retrain cycles—can retain accuracy but becomes infeasible at scale due to high computational cost. One-shot pruning, which removes parameters in a single step without retraining, offers a more scalable alternative but often results in severe accuracy degradation. For instance, pruning 80% of the parameters from RegNetX-32GF (a 100M+ parameter model) drops ImageNet accuracy from 80% to 1%, rendering the model unusable. This talk uncovers a new and fundamental bottleneck behind such failures: signal collapse, a previously overlooked phenomenon that disrupts the network’s ability to distinguish between inputs. To address this, a simple and efficient method called REFLOW is introduced, enabling sparse networks to recover strong performance without retraining or gradient computation. On RegNetX-32GF, REFLOW lifts accuracy from 1% to 73% at 80% sparsity—in under 15 seconds. These findings reframe the challenges of one-shot pruning and open new opportunities for practical and efficient deployment of deep learning models.

PGR Seminar with Leonid Nosovitsky + Xinya Gong

The next PGR seminar is taking place this Friday 11th April at 2PM in JC 1.33a

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

Leonid Nosovitsky

Title: Bridging Theory and Practice: Advancing Multiparty Session Types for Industry Use

Abstract: Multiparty Session Types (MPST) provide a typing discipline for communication protocols. They allow to statically check that a code implementation conforms to a specified protocol; they can also verify that a protocol satisfies many safety properties like liveness and deadlock freedom, which are crucial in concurrent communicating systems.  Despite huge improvements in MPST research, different extensions have limitations.  For example, one of extensions is crash-handling. This branch was motivated by lack of network reliability to make MPST framework more applicable and usable in industrial scenarios. The crash-semantics theory introduced by Barwell et al. does not involve any constraints relaxations, which makes it very intricate for adoption in practical scenarios.  Our project concentrates on addressing usability limitations to make MPST more integrable into industrial applications.

Xinya Gong

Title and Abstract TBC

PGR Seminar with Zihan Zhang + Berné Nortier

The next PGR seminar is taking place this Friday 4th April at 2PM in JC 1.33a

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

Zihan Zhang

Title: FedOptima: Optimizing Resource Utilization in Federated Learning

Abstract: Federated learning (FL) systems facilitate distributed machine learning across a server and multiple devices. However, FL systems have low resource utilization limiting their practical use in the real world. This inefficiency primarily arises from two types of idle time: (i) task dependency between the server and devices, and (ii) stragglers among heterogeneous devices. We propose FedOptima, a resource-optimized FL system designed to simultaneously minimize both types of idle time; existing systems do not eliminate or reduce both at the same time. FedOptima offloads the training of certain layers of a neural network from a device to server using three innovations. First, devices operate independently of each other using asynchronous aggregation to eliminate straggler effects, and independently of the server by utilizing auxiliary networks to minimize idle time caused by task dependency. Second, the server performs centralized training using a task scheduler that ensures balanced contributions from all devices, improving model accuracy. Third, an efficient memory management mechanism on the server increases scalability of the number of participating devices. Four state-of-the-art offloading-based and asynchronous FL methods are chosen as baselines. Experimental results show that compared to the best results of the baselines on convolutional neural networks and transformers on multiple lab-based testbeds, FedOptima (i) achieves higher or comparable accuracy, (ii) accelerates training by 1.9x to 21.8x, (iii) reduces server and device idle time by up to 93.9% and 81.8%, respectively, and (iv) increases throughput by 1.1x to 2.0x.

Berné Nortier

Title: Shortest paths and optimal transport in higher-order systems

Abstract: One of the defining features of complex networks is the connectivity properties that we observe emerging from local interactions. Nevertheless, not all networks describe interactions which are merely pairwise. Recently, different frameworks for modelling non-dyadic, higher-order, interactions have been proposed, garnering much attention. Of these, hypergraphs have emerged as a versatile and powerful tool to model such higher-order networks. However, the connectivity properties of real-world hypergraphs remain largely understudied. A first, data-driven, work introduces a measure to characterise higher-order connectivity and quantify the relevance of non-dyadic ties for efficient shortest paths in a diverse set of empirical networks with and without temporal information. The analysis presents a nuanced picture.

A second work (in progress) considers higher-order simplicial networks within the context of optimal transport, where shortest paths do not always lead to optimal resource allocation. We extend the existing framework to the higher-order setting to explore to what degree this additional degree of freedom influences the flux of resources in a system of interest.

SACHI Seminar, Stavroula Pipyrou – Radical Imagination: Knowledge Through Generations

We are pleased to share our upcoming SACHI seminar this week by Dr Stavroula Pipyrou, a Senior Lecturer in Social Anthropology at the University of St Andrews and Founding Director of the Centre for Minorities Research.

📅 Wednesday 2nd April  | 🕛 13:00 – 14:00 PM | 📍 JCB, Room 1.33A

Title:

Radical Imagination: Knowledge Through Generations

Abstract:

“Today we will engage with insights from a young Greek interlocutor who did not live the Cold War period firsthand. She relates to the legacies of the Cold War through radical imagination, projecting that it is only logical that the affects of the era have left irreversible psychological marks on the people who experienced it. The talk proposes a theory of psychic time and generational battles for belonging. There is a critique of history as taught in school textbooks when compared to the lived experiences of history in the present.”

Bio:

Stavroula is Senior Lecturer in Social Anthropology at the University of St Andrews, Founding Director of the interdisciplinary Centre for Minorities Research and editor of the interdisciplinary book series Routledge Advances in Minority Studies. She works on minority politics, displacement, governance, and the Cold War. She is the author of “The Grecanici of Southern Italy: Governance, Violence, and Minority Politics” (University of Pennsylvania Press, 2016) and “Lurking Cold War: Life Through Historical Communion” (Berghahn 2025). Lurking Cold War explores the entangled registers of the Cold War that continue to stalk the social landscape in Italy and Greece. Critiquing the connections between global categories and individual experiences, Lurking foregrounds Cold War resonances through materiality, imagination, speculation and affect, in literature, bureaucracy and in the family. A theory of methexi illustrates how people and history are brought into communion, blurring the boundaries between known and unknown, reality and imagination, and form and interpretation. The result is an articulation of history that matters in a way that matters.

PGR Seminar with Joe Loughney

The next PGR seminar is taking place this Friday 28th March at 2PM in JC 1.33a

Below are the Title and Abstract for Joe’s talk – Please do come along if you are able.

Title: Flexible-Order Symmetry Breaking in the Subgraph Isomorphism Problem

Abstract: The Subgraph Isomorphism Problem has many applications, including bioinformatics, computer vision and graph databases. Current state-of-the-art solvers using constraints programming techniques can handle cases with up to 1000 pattern vertices and 10,000 target vertices. We explore various approaches to variable and value symmetry breaking in the problem (and viable strategies to combine the two), implemented in the Glasgow Subgraph Solver, and introduce the notion of ‘flexible ordering’ on symmetry breaking constraints.

SACHI Seminar, Mark Zarb – Bridging Minds and Machines: Redefining Computing Education

We are pleased to share our upcoming SACHI seminar by Dr Mark Zarb, an Associate Professor based within the School of Computing, Engineering and Technology at RGU:

📅 26th March | 🕛 13:00 – 14:00 PM | 📍 JCB, Room 1.33A

Title:

Bridging Minds and Machines: Redefining Computing Education

Abstract:

Since 2009, Dr Zarb has been exploring the evolving landscape of pedagogical research, collecting ideas from across disciplines and trends. In this acronym-filled talk, he offers a guided tour through some of the latest research at RGU — from grappling with the ethical dilemmas posed by conversational AI in education, to exploring “shadow podcasts” as informal learning tools. We will look at practical challenges, unexpected questions and at how rapidly shifting technology continues to shape how (and why) we teach and learn.

Bio

Dr Mark Zarb is an Associate Professor based within the School of Computing, Engineering and Technology at RGU

His main research focus is within computing education, having led international working groups on transitions into higher education in 2018 and post-pandemic educational landscapes in 2021 and 2022.

He received his PhD (2014, University of Dundee) for work exploring the role of verbal communication styles in pair programming. His various roles and experiences allow him a wide and international perspective on computing education.

PGR Seminar by Constantine Theocharis + Yigit Yazicilar

The next PGR seminar is taking place this Friday 21st March at 2PM in JC 1.33a

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

Constantine Theocharis

Title: Efficient Programs with Dependent Types

Abstract:

Dependent types allow us to program using the full power of set theory at our disposal. We can encode conditions of arbitrary complexity, and then show that these conditions are met by our programs, statically. While this paradigm is very effective for verifying systems, often their real-world implementations are done in languages without these verification capabilities, because they produce more efficient programs. In this talk, I will explore some of the main sources of inefficiency in (functional) languages with dependent types, and some work that aims to mitigate these, so that verification and implementation can happen in the same language. A common pattern in these languages is to have ‘refinements’ of data which carry along with them proofs of the properties we care about. The first piece of work is about how to make these refinements true zero-cost abstractions. Another source of inefficiency is that these languages must heap allocate almost everything since the sizes of types cannot always be known at compile time. The second piece of work is about how to keep track of type sizes as part of the type system, so that all heap allocations are explicit and unnecessary for the most part.

Yigit Yazicilar

Title: Automated Nogood-Filtered Fine-Grained Streamlining

Abstract:
We present an automated method to enhance constraint models through fine-grained streamlining, leveraging nogood information from learning solvers. This approach reformulates the streamlining process by filtering streamliners based on nogood data from the SAT solver CaDiCaL. Our method generates candidate streamliners from high-level Essence specifications, constructs a streamliner portfolio using Monte Carlo Tree Search, and applies these to unseen problem instances. The key innovation lies in utilising learnt clauses to guide streamliner filtering, effectively reformulating the original model to focus on areas of high search activity. We demonstrate our approach on the Covering Array Problem, achieving significant speedup compared to the state-of-the-art coarse-grained method. This work not only enhances solver efficiency but also provides new insights into automated model reformulation, with potential applications across a wide range of constraint satisfaction problems.

SACHI Seminar with Aluna Everitt – Democratising the Design and Development of Emerging Technologies

We are pleased to share our upcoming SACHI research seminar by Dr Aluna Everitt, a lecturer in the Department of Computer Science and Software Engineering at the University of Canterbury, New Zealand:

📅 Today | 🕛 12:00 – 1:00 PM | 📍 JCB, Room 1.33B

Title:

Democratising the Design and Development of Emerging Technologies

Abstract:

My research focuses on democratising the development of emerging technologies. More specifically, by establishing accessible approaches for designing and building emerging technologies such as robotics, wearables, and shape-changing interfaces. To advance the field, my research focuses not only on understanding these technologies (e.g., their design), but also how to build them (e.g., engineer them), and how to innovate with them (e.g., application). In this talk, I will go into detail about some of the projects I have worked on around this topic across the fields of HCI, Design, and Engineering.

Bio:

Dr. Aluna Everitt is a lecturer in the Department of Computer Science and Software Engineering at the University of Canterbury, New Zealand. Prior to moving to Christchurch (NZ), she was a Research Associate in the Cyber-Physical Systems group at the University of Oxford and a Junior Research Fellow at Kellogg College, University of Oxford. She was also a Senior Visiting Researcher and postdoc at the University of Bristol (BIG Lab). Dr. Everitt was awarded her PhD in Computer Science from Lancaster University, specializing in Human-Computer Interaction (HCI). As a multi-disciplinary researcher, her areas of interest and expertise lie across the fields of HCI, Design, and Engineering. She has a particular interest in conducting both quantitative and qualitative research which combines a mix of engineering fabrication approaches for iterative prototyping, together with collaborative design (co-design) to encourage users and experts from different domains to develop content and applications for the next generation of interactive hardware systems and interfaces (e.g., shape-changing displays, wearables, and robotics).

PGR Seminar with Mirza Hossain

The next PGR seminar is taking place this Friday 14th March at 2PM in JC 1.33a

Below are the Title and Abstract for Mirza’s talk – Please do come along if you are able.

Title: BioFuse: Optimizing Biomedical Embeddings with Foundation Models

Abstract: Pre-trained foundation models have revolutionized biomedical AI, excelling in specialized domains like radiology and histopathology. However, integrating multiple models remains a challenge due to compatibility and feature fusion issues. BioFuse is an open-source framework designed to optimize biomedical embeddings by automatically selecting and fusing the best model combinations. Leveraging 9 state-of-the-art foundation models and a grid search strategy, BioFuse generates task-specific embeddings that improve downstream classification. On the MedMNIST+ benchmark, it achieves SOTA AUC in 5/12 datasets while maintaining near-SOTA performance in others. Surprisingly, our experiments reveal strong cross-modal capabilities, where models trained on one modality perform well on others. With a high-level API and an extensible architecture, BioFuse streamlines model integration and paves the way for new insights in biomedical data fusion.

PGR Seminar with Ben Claydon and Erdem Kus

The next PGR seminar is taking place this Friday 28th February at 2PM in JC 1.33a

Below are Titles and Abstracts for Ben and Erdem’s talks – Please do come along if you are able.

Ben Claydon

Title: Mechanisms for Similarity Search

Abstract:

Similarity search encompasses the task of finding those objects in a large collection which are most alike to, in some way, an object presented by the user as a query. The domain of these objects is wide, from images to text to chemical structures. This task becomes yet harder when the database becomes extremely large, and a sublinear query time with respect to the database size becomes a requirement. This talk discusses why the problem becomes so hard when presented with complex data, and how algorithms and data structures can be engineered to serve these queries.

Erdem Kus

Title: Frugal Algorithm Selection

Abstract: When solving decision and optimisation problems, many competing algorithms (model and solver choices) have complementary strengths. Typically, there is no single algorithm that works well for all instances of a problem. Automated algorithm selection has been shown to work very well for choosing a suitable algorithm for a given instance. However, the cost of training can be prohibitively large due to running candidate algorithms on a representative set of training instances. In this work, we explore reducing this cost by choosing a subset of the training instances on which to train. We approach this problem in three ways: using active learning to decide based on prediction uncertainty, augmenting the algorithm predictors with a timeout predictor, and collecting training data using a progressively increasing timeout. We evaluate combinations of these approaches on six datasets from ASLib and present the reduction in labelling cost achieved by each option.