Dynamic Infrastructure and University of Bath Test How AI Approaches Perform on Real-World Infrastructure

A joint research study compares different AI approaches and tests how well they generalise to infrastructure they have never seen before.

NEW YORK and BATH, England, Sept. 30, 2026 /PRNewswire/ -- Dynamic Infrastructure, the Engineering AI platform for critical civil assets, today confirmed an ongoing research collaboration with the University of Bath's Department of Architecture and Civil Engineering. Working with Thomas Kjeldsen, Professor of Hydrology and Water Engineering, the project is being carried out as part of Arline Osorio Moreno's MSc research and explores a practical question with major implications for infrastructure AI: how well different AI approaches hold up on a real engineering task, especially when applied to assets and sites they have never seen before.

The study examines how computer vision can identify infrastructure conditions from inspection images, focusing on blockage and obstruction in drainage assets. A central part of the work is cross-site generalisation: a model that performs well on the sites it was trained on may perform very differently on a network in another region, where construction, camera equipment and inspection practices vary. That gap matters enormously to any agency deciding whether an AI system is reliable enough for real-world use.

To test this, the research compares several families of approach against the same task: vision-language models, vision foundation models, supervised deep learning and classical machine-learning methods. Each carries different trade-offs in accuracy, robustness, how much labelled data it needs and how well it transfers between sites. Early results show strong and meaningful differences between both methods and sites. They also indicate that current frontier vision-language models, despite their broad capabilities, do not yet provide the robustness required for reliable real-world implementation on this specialised task, while more targeted approaches can perform substantially better.

"What stood out to us is how differently the models behave on the same engineering task," said Arik Voronov, AI R&D Lead at Dynamic Infrastructure. "The biggest, most general models are not always the strongest choice here, and more specialised approaches can be much more consistent. The other important question is whether that performance holds up when the model is exposed to different sites, inspection conditions and asset characteristics. That combination — model choice and robustness across real-world variation — is where the interesting picture starts to emerge."

"This project demonstrates the important role that partnerships between academia and industry can play in advancing engineering innovation," said Professor Kjeldsen. "By combining the University of Bath's research expertise with Dynamic Infrastructure's practical knowledge, we are able to test emerging AI technologies against pressing infrastructure challenges. Collaborations like this help ensure that new AI tools are not only scientifically rigorous but also capable of delivering tangible benefits for infrastructure operators and wider society."

This has been an exciting collaboration and the partners intend to continue working together with the aim of developing the work into a peer-reviewed publication.

About Dynamic Infrastructure: Dynamic Infrastructure is the market-leading Engineering AI agent platform and system of record for critical civil assets, managing thousands of structures across 15 states and countries. The platform delivers autonomous risk intelligence to predict deterioration and assess structural health, replacing subjective opinions with data-driven clarity to help prevent catastrophic failures. Led by deep-tech entrepreneurs and industry experts, the company scales engineering judgment to address global budget constraints, the workforce crisis, and the aging infrastructure cliff. For more information, please visit https://diglobal.tech/

About the University of Bath:  

The University of Bath is one of the UK's leading universities, recognised for high-impact research, excellence in education, an outstanding student experience and strong graduate prospects. 

• We are ranked among the top 10% of universities globally, placing 125th in the QS World University Rankings 2027. 

• We are ranked in the top 10 in the Complete University Guide, Guardian University Guide and The Times and The Sunday Times Good University Guide. 

• We are ranked No. 1 in the UK for happiest graduates in The Guardian University Guide. 

• The University achieved a triple Gold award in the last Teaching Excellence Framework 2023, the highest awards possible, for both the overall assessment and for student outcomes and student experience. The Teaching Excellence Framework (TEF) is a national scheme run by the Office for Students (OfS). 

Research at Bath is shaping a better future through innovation in sustainability, health, and digital technologies. Find out all about our Research with Impact: http://bit.ly/3ISz1Wu  

Contact

Leah Stern, Dynamic Infrastructure Media Desk
[email protected] / +44-747-019-6826

SOURCE Dynamic Infrastructure