Plenary Lecture 2

Room: Audimax
Time: Monday, 28. September 2026, 10:20 - 11:00
Chair: Michael Beer

Lizzy Cross

Professor
Dynamics Research Group, University of Sheffield, UK

Into the deep; can physics keep us green?

Large language models and AI use have become a normal part of everyday life for many of us (and if not for you, then certainly for your students). The future of engineering practice is also likely to be one where AI design, analysis, and reporting plays a frequent role. As engineers and experts, we now need to anticipate the benefits, threats and unintended consequences of AI use professionally, particularly where tools used are generic and not designed with engineering applications in mind. In the generic case, concerns on reliability, interpretability, trustworthiness come first to mind. However, this talk argues that we should also consider and mitigate the environmental impact of AI when interwoven into engineering workstreams.

Environmental concerns around the accelerated use of AI are well documented, whether considering energy draw or water reliance for data centres. So far, however, there has been little consideration of how this will accelerate further as we start using these tools in our professional context, particularly where we might expect considerable interplay between AI and other approaches with a heavy computational burden (e.g. FEA, CFD). With the main AI providers not currently required to share their energy &/water use, barriers to quantifying and projecting this impact are large. However, regardless, there are steps we can take as engineers to mitigate the potential impact meanwhile (the foremost putting processes in place to ensure that no nut cracking is attempted with AI sledge hammers).

This talk will specifically consider how embedding physics directly into machine learning inference (which we will consider as the AI component) could help to lower the environmental burden from compute - in this case for examples in structural dynamics. We will explore a number of avenues where our knowledge embedded in simple ways can reduce impact, but also highlight where the reverse might be true.

Short Bio

Lizzy Cross is a Professor in the Dynamics Research Group at the University of Sheffield, with research interests spanning the fields of structural health monitoring (SHM), machine learning and nonlinear system identification. Most of her research projects focus on the analysis of large datasets from monitored structures, where she employs data-driven algorithms to extract useful information, however, she has recently completed an EPSRC Innovation Fellowship pioneering a physics-informed machine learning approach for these problems. Lizzy is a co-director of the Laboratory for Verification and Validation, a state-of-the-art dynamic testing facility (lvv.ac.uk). She has published over 170 articles, including 50+ journal papers and 8 invited book chapters. She serves as an Associate Editor for the Cambridge Press Data-centric Engineering journal, as well as for the open access journal of Structural Dynamics. She awarded the Achenbach medal which recognises an individual (within 10 years of PhD) who has made an outstanding contribution to the advancement of the field of SHM in 2019.