The next PGR seminar is taking place this Friday 30th May at 2PM in JC 1.33a

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

Lina Hadji-Kyriacou

Title: Context-PEFT: Efficient Cross-Domain Transfer Learning

Abstract: Parameter-Efficient Fine-Tuning (PEFT) techniques such as LoRA, BitFit and IA3 have demonstrated comparable performance to full fine-tuning of pre-trained models for specific downstream tasks, all while demanding significantly fewer trainable parameters and reduced GPU memory consumption. However, in the context of cross-domain transfer learning, the need for architectural modifications or full fine-tuning often becomes apparent. To address this we propose Context-PEFT, which learns different groups of adaptor parameters based on the current input domain. This approach enables LoRA-like weight injection without requiring additional architectural changes. Our method is evaluated on the COCO captioning task, where it outperforms full fine-tuning under similar data constraints while simultaneously offering a substantially more parameter-efficient and computationally economical solution.

Victor Yuan

Title: Methodologies for Creating Interactive and Lifelike Historical Characters Based on MetaHuman

Abstract: Virtual characters have long held promise as pedagogical tools in heritage education, particularly for creating immersive interactions with historical figures. Researchers have envisioned systems capable of emulating these figures, enabling users to engage in life-like, face-to-face dialogues over time. While technological constraints historically limited such applications, recent advancements in computational graphics and language models have now made them viable. This paper presents a framework for constructing interactive virtual character systems, outlining their core components through two critical dimensions: photorealism and interaction. The photorealism dimension leverages modern graphics tools to achieve high-fidelity visual representation, while the interaction dimension utilizes language models to enable socially believable and contextually responsive dialogue. We examine the necessity of each component and analyze available technological solutions with their respective trade-offs. Beyond the technical framework, we discuss potential future improvements and address ethical and practical concerns inherent to such systems. By synthesizing current technologies and their applicability, this work provides institutions with practical guidance for developing customized interactive systems that balance functionality with cost-efficiency.

The next PGR seminar is taking place this Friday 23rd May at 2PM in JC 1.33a

Below are the Titles and Abstracts for Thomas and Charis’ talks – Please do come along if you are able.

Thomas Martin

Title: From Isolated to Continuous Automated Sign Language Recognition

Abstract: Sign languages are full-fledged visual natural languages combining manual and non-manual features. As the name suggests, Automated Sign Language Understanding (ASLU) aims to automate tasks involving sign language. A primary obstacle to ASLU is the creation of appropriate datasets. Indeed, most datasets focus on materials gathered from TV broadcasts covering limited topics, which fail to accurately reflect sign language in the wild. Moreover, annotating such datasets is a prohibitively costly process. With the end goal of Sign Language Translation (written/spoken language to sign language) in mind, ASLU research has transitioned from Isolated to Continuous Sign Language Recognition. However, sign language intricacies have made this transition non-trivial.

Charis Hanna

Title: Enhancing Deep Learning Approaches for the Automated Monitoring of Dense Seabird Colonies

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 semi-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.

The next PGR seminar is taking place this Friday 16th May at 2PM in JC 1.33a

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

Tilcia Woodville-price

Title: Definitely, Maybe? Communicating Uncertainty In Medicine

Abstract: An inherent aspect of healthcare and data science alike, uncertainty is present at every step of data utilisation, including its collection, analysis and dissemination. Accounting for and disclosing uncertainty is an ongoing challenge faced by many disciplines. In medicine, effective communication of uncertainty is essential for shared decision-making, with important considerations surrounding the user (patients, clinicians, policymakers), the underlying information, and the means of communicating it (data visualisation). Current medical research has focused on risk communication, often failing to evaluate more complex aspects of uncertainty. In contrast, the information visualisation community has more widely researched uncertainty visualisation in other domains, but insights remain limited regarding best practice. This interdisciplinary research aims to empirically assess different forms of visualising uncertainty in medicine, evaluate how user characteristics influence comprehension, and explore new forms of communicating it through data-driven storytelling.

Thu Nguyen

Title: Multimorbidity Dynamics in Scotland: Health inequality and Trajectories in chronic disease accrual and mortality across the lifespan

Abstract: Multimorbidity – the co-occurrence of two or more chronic diseases – is a growing global concern, and is associated with higher risk of mortality, worse quality of life and substantial financial burden. About one third of the world’s population has multimorbidity. Care for multimorbid patients in the UK accounts for more than 55% of NHS costs and 75% of primary care prescription costs. The focus of multimorbidity epidemiology so far has largely been on the static clustering of diseases through cross-sectional analyses, with less emphasis on the trajectories of disease onset and the sequence in which conditions develop. Understanding the order in which diseases occur and its impact on patient outcomes can help identify high-risk trajectories and aid healthcare resource planning by identifying target populations for preventive interventions, ultimately leading to earlier diagnosis and management. Using linked electronic health records (EHRs) on 858789 individuals (2005-2021), this study aims to employ multistate modelling to explore the dynamics of multimorbidity trajectories, measure chronic disease accrual, incorporating social factors to unveil the health inequality in Scotland.