Skip to main content

IDSAI Funding for Pilot Projects Announced

The IDSAI is delighted to announce that six new research projects have been selected to receive an IDSAI Research Award or Exeter-Turing Pilot Research grant from the latest round of applications. These six projects start in July 2021. 

IDSAI Research Awards and Turing-Exeter Pilot Research Grants aim to support early-phase development of ideas for innovative, interdisciplinary data science and AI research across Exeter.  Academic members of staff from all Colleges were invited to apply for the time of an IDSAI Research Fellow to work with them on a pilot research project.  You can read about each project below.

For further information about any of the projects please contact:

COLLABORATORS:David Reynolds (CGES), Paul Halloran (Geography), Bryan Black (University of Arizona)
IDSAI Research Fellow: George De Ath

Description: Growth rings formed in trees, clam shells, corals and fish provide unique records of the environmental conditions at the time the ring was formed. Given the great age that these species can reach, these growth ring records can provide long-term perspectives on modern climate variability. These records are ultimately used to help improve the accuracy of climate models used for predicting future climate change. However, the process of developing these records is painfully slow as each ring is manually measured and validated using statistical techniques. Whilst advances have been made in the software used to perform these analyses, significant potential remains for optimising the methodology and improving our understanding of uncertainties within these data. This project seeks to explore the potential of using automated image analysis techniques and Bayesian statistics to automate the analysis of growth rings formed in trees and clams. It is hoped that these advanced techniques will speed up the development of new long-term environmental records whilst preserving the statistical rigour with which these records are built. These new and extended records will help further our understanding of how the oceans and atmosphere change through time and ultimately improve the accuracy of future climate forecasts.  
COLLABORATORS: Dr Vicki Gold (CLES) and Dr Danielle Paul (Turing Fellow, University of Bristol)
IDSAI Research Fellow: Dmitry Kangin
Description: Cryo-electron tomography (cryoET) is a specialised 3D high-resolution imaging method used to visualise protein complexes and determine their structures. We image protein filaments, which are long hair-like assemblies that play important roles in a plethora of cellular processes. Examples include actin and myosin found in muscle, as well as pilins/flagellins which drive microbial motility.  By averaging many thousands of copies of protein filaments together from our 3D data, we can determine high-resolution protein structures at atomic-level detail (Fig. 1). Such structures help to reveal fundamental insight into their function, which provides key knowledge for healthcare and drug discovery programmes. A major limitation of our work is filament identification and selection for downstream processing, which is mostly a manual process. In this proposal, we aim to establish the feasibility of automated filament identification in cryoET data using two- and three-dimensional convolutional neural networks. Removing this restrictive bottleneck will change the way in which cryoET imaging data can be processed, increasing efficiency and increasing the accessibility of our method to the scientific community.
COLLABORATORS: Dr Chico Camargo, CEMPS, University of Exeter, affiliated academic at IDSAI, Co-I: Dr Barbara McGillivray, Turing Research Fellow, The Alan Turing Institute, Senior Research
IDSAI Research Fellow: Dr Bertrand Nortier
Description: Language is an important force shaping politics and guiding collective sensemaking. This is evident with the rise of terms such as "gender pay gap", "Brexit" even "lockdown". In this context, lexical semantic change, i.e. the change in meaning of words, is fast, wide, and weaponised. This makes it pressing to ask: can we detect and measure how meanings change or persist in certain communities? Can we track lexical semantic change at scale, in real time?

This project will develop new data science methods to track the semantic change of political terms. It will combine tools from natural language processing, machine learning, and network science, to make "socially aware" computational models of lexical semantic change. We will do so by extending methods of semantic change detection based on word embeddings, as well as token and sense embeddings, applying them to a series of political terms present in debates in the UK Parliament over the last decade, and incorporating metadata on party affiliation, debate topic, and debate participants, to reveal how word meanings were shared – and built – across different parties and debates. As such, this project will enable us to measure the dynamics of semantic divergence and convergence, as well as coordinated behaviour in the UK Parliament.
COLLABORATORS:  Jude Meakin (Biomedical Physics), Jonathan Fulford (Medical Imaging), Karen Knapp (Medical Imaging), Greg Slabaugh (Computer Science, QMUL)
IDSAI Research Fellow: Dr Bertrand Nortier
Description:  Our project aims to develop computer software that can automatically assess spinal muscle size and quality from medical images. Our spine muscles are essential for controlling the movement of our upper body, allowing us to stand upright, adopt a range of postures, and lift objects from the ground. These muscles, however, are often smaller and fattier in people who have conditions such as back pain and osteoporosis. Smaller and fattier muscles indicate that the muscle has become weaker, making it less able to control the spine. However, we do not know whether muscle weakness is a symptom or a cause of these conditions. Large scale studies are required to answer this question, and these studies need suitable methods for quickly and accurately assessing muscle size and quality. Therefore, our project aims to develop computer software that can identify the spinal muscles in magnetic resonance images and then automatically calculate the muscle size and the amount and distribution of the fat within the muscle. Once we have developed our software, we will use it on images we already obtained to determine how muscle size and fat varies in people with and without osteoporosis.
COLLABORATORS:  Dr James Hickey (CSM), Prof. Hywel Williams (Alan Turing Fellow & Comp. Sci.) & Dr Rudy Arthur (Comp. Sci.); Co-I = Michelle Spruce (Comp. Sci.); External Collaborator: USGS
IDSAI Research Fellow: Dr Ravi Pandit
Description: Volcanic crises cause significant damage, casualties and economic loss. Social sensing is the systematic analysis of unsolicited social media data to observe real world events. Our project aim is to use social sensing to analyse the social and economic impact of volcanic crises through space and time, and the change in emotional response to a crisis as it unfolds, from pre-eruption warnings through to post-eruption recovery. Current understanding of social impacts of volcanoes are poorly understood and there is no widely-accepted best practice in communicating volcano hazard information; our proposed analyses will contribute to solving these problems. We will initially focus on the 2018 Kilauea eruption in Hawaii, USA, in collaboration with the United States Geological Survey (USGS), before looking at volcanic crises from a more global standpoint. By developing a set of tools to socially sense the economic and social impact of volcanic hazards, and societal reaction to official hazard management communications, we will help define best-practise in this area to benefit the 10% of the world’s population living with volcanic hazards.
COLLABORATORS:  Dr Miriam Koschate-Reis, Senior Lecturer Social Psychology, Psychology (CLES), Dr Travis Coan, Senior Lecturer Politics, Politics (CSSIS), Prof Susan Banducci, Politics (CSSIS) and Turing Fellow, Postdoctoral collaborators: Dr Elahe Naserian-Hanzaei, Politics (CSSIS), Ms Alicia Cork, Psychology (Bath)
IDSAI Research Fellow: Ravi Pandit

Description: One of the most fundamental questions in leadership research is whether leaders shape groups (identity entrepreneurship), or whether changes to a group’s self-understanding require a change in leadership (leader prototypicality). The recent Trump presidency illustrates this question well: On the one hand, political analysts have suggested that Trump was elected leader in response to shifts in the Republican Party created by the Tea Party movement. On the other hand, it has been proposed that Trump has shaped “American Nationalism” as a movement and embedded it in the Republican party. However, to the best of our knowledge, there is no research examining prototypical leadership and identity entrepreneurship together over time by examining changes to a group’s self-understanding, and the fit of leaders to this self-understanding. This is a particularly important question given the rise of extreme right-wing groups across the world, and their public endorsement by “mainstream” political leaders. 

Our recent research suggests that a group’s self-understanding is reflected in the linguistic style of its members. By analysing online forum data of right-wing groups and speeches of members of the U.S. congress (leaders), this project aims to examine the self-understanding of a wide spectrum of right-wing and conservative political groups in the U.S., and the role that leadership plays in this self-understanding. Combining computational and statistical methods with political and psychological theory will allow us to gain novel insights into (1.) the dominant group types found in right-wing U.S. politics, and whether there is a differentiation in self-understanding between mainstream and right-wing groups, (2.) the extent to which leaders drive changes in the self-understanding of their group, or whether changes in the group’s self-understanding require a change in leadership.