PhD Studentship: IMPACT-RISE: Infrastructural Surrogate Modelling Using Physics-informed and Interpretable Machine Learning for Community Resiliency and Sustainability Evaluation

Location:

Department of Computer Science, Streatham Campus, Exeter,

The Department of Computer Science at the University of Exeter is currently accepting applications for a fully funded PhD studentship, with a negotiable enrolment date open until January 2026 or earlier. For eligible students the studentship will cover Home or International tuition fees plus an annual tax-free stipend of at least £18,622 for 4 years full-time, or pro rata for part-time study. 

Project Description

The IMPACT-RISE project is a pioneering initiative that seeks to revolutionize the field of community resiliency and sustainability analysis through a machine learning (ML) and explainable artificial intelligence (XAI) outlook. The project marks a significant advancement in improving public safety against both low-probability high-impact events and high-probability events with long-term impacts. It focuses on the development of state-of-the-art infrastructural surrogate models using physics-informed and interpretable ML techniques. Our aim is to comprehensively analyse and mitigate the risks posed by diverse extreme events, both natural and anthropogenic (including earthquakes, floods, storms, climate change), on built environment. The primary goal is to enhance our understanding and predictive capabilities, thereby improving decision-making processes to effectively reduce the impact of these hazards on infrastructure systems.

Central to IMPACT-RISE project is the development of data-driven deep learning (DL) based surrogate models that simulate the complex behaviours of infrastructure systems under conditions posed by various hazards (occurring independently and concurrently). These models will be trained while appropriately infusing physics (such as structural dynamics), ensuring not only high accuracy but also enhanced interpretability – a crucial factor for decision-makers in risk management and emergency response. To further boost the interpretability of the DL based surrogate models, principles of explainable artificial intelligence (XAI) will be integrated for a deeper understanding of the models' decision-making processes. Working on the project involves the meticulous collection, development, and analysis of diverse infrastructural and hazard related data sets, ranging from historical incident records to real-time infrastructural sensor data, community maps, and more. Furthermore, the project requires augmentation of real recorded data with simulation data obtained through structural finite-element modelling and analyses.

IMPACT-RISE project aims to provide accurate, reliable, and accessible models, thereby playing a pivotal role in fortifying community resilience and sustainability against various hazards. These innovative tools will be instrumental in pinpointing vulnerabilities, optimizing resource distribution, and crafting effective emergency response plans. IMPACT-RISE is grounded in collaborative effort, bringing together a diverse team of specialists in machine learning, civil engineering, and risk analysis. We are committed to align our models with the practical realities and unique challenges of different communities. Through this integrated and cooperative approach, IMPACT-RISE is set to establish new standards in community protection and infrastructure resilience, confronting the diverse challenges of the 21st century with advanced technological solutions and strategic insights.

The project is open-ended and offers flexibility, inviting applicants to suggest their unique ideas that align with the overarching theme and objectives of the initiative.

Advert information