The Cambridge Graphene Centre (CGC) was delighted to host Prof. Paulette Clancy (Johns Hopkins University, USA) for a special QAMSS SRI Advanced Technology Lecture as part of the Lent Term 2026 programme.
Prof. Clancy’s lecture, “Bayesian Optimization-Driven Materials Discovery and Design: Hype, Power and Challenges,” explored how Bayesian optimisation is increasingly being used to accelerate materials discovery and design by enabling efficient exploration of complex design spaces under limited experimental or computational budgets. She provided an accessible overview of why these methods have gained traction across materials research and how they can help prioritise experiments or simulations in a principled, data-driven manner.
A key focus of the talk was a realistic and balanced view of both the strengths and the limitations of Bayesian optimisation in practice. Prof. Clancy discussed where the approach can be particularly powerful, and highlighted common challenges that can hinder performance when applied to real materials problems. Topics included the critical role of data quality, how uncertainty quantification shapes decision-making, and the importance of selecting appropriate objective functions to align optimisation outcomes with scientific and engineering goals. The lecture also drew on representative examples and practical lessons learned from recent work.
The event prompted an engaging discussion with the QAMSS audience, reflecting strong interdisciplinary interest at the interface of materials science, computation, and machine learning. We thank Prof. Clancy for an excellent lecture and for her visit to CGC.
