Luke Dicks is a research scientist based at IBM Research Europe (UK). His interests lie in the adaptation and application of physics methods to AI and machine learning algorithms, with a focus on developing tools to provide greater explainability, reproducibility, and transferable understanding between different models.
Luke received his PhD in Theoretical Chemistry from the University of Cambridge. His early research focussed on enhanced sampling methods for simulating biomolecules and their aggregation. He then adapted these enhanced sampling methods to explore and map the solution spaces of machine learning models, specifically focussing on clustering algorithms. After his PhD he was awarded a knowledge transfer fellowship by the EPSRC and worked between the University of Cambridge and IBM Research Europe (UK). In this fellowship he extended the novel methods to additional machine learning models to provide novel and transferable understanding between them.