School of Mathematical Sciences

AI-Driven infrastructure inspection: continual improvement through reinforcement learning

Project description

This project is an exciting opportunity to undertake industrially linked research in partnership with the Manufacturing Technology Centre (MTC). It is based within the School of Mathematical Sciences at the Faculty of Science, University of Nottingham, which amongst its wide research portfolio, conducts cutting edge research into the development novel causal inference, automation and robotic control algorithms.

This is 3-year fully funded studentship and is only open to UK home students. The successful applicant will receive a generous tax-free annual stipend of £25,000 plus payment of their full-time home tuition fees. Additionally, £2,000 per annum is provided for consumables, travel, etc. Due to funding restrictions this PhD position is only available to UK nationals. As this position is sponsored by the MTC, any successful candidate would need to pass the sponsors own security checks prior to the commencement of the PhD.

Vision

We are seeking for a highly motivated PhD student to conduct cutting edge research of the AI techniques and reinforcement learning, a technology which has powered many of the recent groundbreaking self-guided game engines and large language models.

Together we will study how the existing and emerging paradigms in reinforcement learning can be utilized to power automated annotation and diagnostic software of critical infrastructure via continual learning from sensor feedback.

Motivation

The rapid evolution of technologies in manufacturing, construction monitoring infrastructure integrity presents significant challenges in adaptability and efficiency. These industries are characterised by dynamic environments where new materials, processes, and systems are continually introduced. Traditional static defect and anomaly detection methods struggle to keep pace with these changes, leading to potential safety risks, operational inefficiencies, and increased costs due to undetected anomalies or false alarms.

Continual learning offers an adaptive solution, enabling defect and anomaly detection systems to evolve dynamically alongside their environments. By employing life-long learning AI, these systems can incrementally learn from new data, adapt to emerging types of anomalies, and maintain high detection accuracy over time. For instance, in manufacturing, continual learning can help identify defects in real-time despite variations in production processes. In construction, it can monitor structural integrity as buildings age or undergo modifications. In healthcare monitoring infrastructure, it can adapt to individual patient changes, ensuring timely detection of critical health events. This PhD will investigate life-long learning AI approaches for automation and control in manufacturing, construction, and healthcare monitoring infrastructure integrity. The research aims to develop flexible, efficient systems capable of continual adaptation, thereby enhancing safety, improving operational efficiency, and contributing to intelligent automation across these vital sectors.

Aim

This PhD aims to develop novel algorithms for Artificial Intelligence (AI) driven continual learning via reinforcement learning (RL), incorporating mechanistic knowledge through causal inference constraints. These algorithms will enable adaptive digital systems for assisted and automated annotation software, as well as diagnostics software, particularly in the contexts of manufacturing and precision imaging. By leveraging causal insights, the project will enhance the systems' ability to learn dynamically from sensor feedback while maintaining consistency and reliability. In the later stages, the research will apply these advancements to a case study on adaptive disassembly lines, demonstrating how continual learning can drive more efficient and sustainable solutions in complex, evolving environments.

You will have the opportunity to join a multidisciplinary team of supervisors: experts in engineering and biochemistry related to different battery technologies; experts in foundational computer science and mathematical foundations of AI; and experts in the industrial utilisation of emerging AI technologies for various manufacturing and built environment inspection processes. 

You will work alongside a team of research engineers based at the MTC, as well as a vibrant cohort of PhD candidates from the School of Mathematical Sciences at the University of Nottingham, the Horizon DTC, and the AI DTC.

Entry requirements

An enthusiastic, self-motivated, PhD candidate with an aptitude for programming and problem solving. The ideal candidate would have (i.e. or expect to have by the start date) a 1st or a 2:1 degree in a STEM field such as Mathematics, Computer Science, Engineering, Physics and others. Prerequisite background in AI or robotics systems would be advantageous but it is not expected. However, we do expect candidates to have adequate experience in coding in at least one object-oriented language (Python, MATLAB, R, C++ etc.). The University actively supports equality, diversity and inclusion and encourages applications from all sections of society. University of Nottingham prides itself on its inclusiveness and close relationship between staff and students with the first Athena SWAN Gold Award in the UK. The Faculty of Engineering provides a thriving working environment for all PGRs creating a strong sense of community across research disciplines. Community and research culture is important to our PGRs and the FoE support this by working closely with our Postgraduate Research Society (PGES) and our PGR Research Group Reps to enhance the research environment for PGRs. PGRs benefit from training through the Researcher Academy’s Training Programme, those based within the Faculty of Engineering have access to bespoke courses developed for Engineering PGRs: including sessions on paper writing, networking and career development after the PhD. The University of Nottingham has outstanding facilities which include £11 million Robotics Lab (The Cobot Maker Space), state-of-the-art high performance computing facilities, as well as some of the best sports facilities in the country.

The MTC is an independent Research and Technology Organisation (RTO) with the objective of bridging the gap between academia and industry and therefore creating a positive, significant impact on society. At the MTC you’ll find mould-breakers and decision makers, creators and doers – all working together to grow our business, nurture our organisation and impact society, every day and everywhere. For more information please visit the MTC website.

Application process

Formal applications are to be made via the MyNottingham system stating the supervisor name and project title. Recruitment of candidates will happen on rolling basis before the application deadline in 4 week windows, so we encourage formal and informal applications to be made as early as possible.

Enquiries to be directed to

Informal inquiries, with a detailed CV and academic transcripts, should be sent to Dr Yordan Raykov (Yordan.Raykov@nottingham.ac.uk) and Dr. Yazan Qarout (Yazan.Qarout@the-mtc.org).

Supervisor contacts

Yordan Raykov
 

Related research centre or theme

 
 

 

 

More information

Full details of our Maths PhD

How to apply to the University of Nottingham

School of Mathematical Sciences

The University of Nottingham
University Park
Nottingham, NG7 2RD

For all enquiries please visit:
www.nottingham.ac.uk/enquire