My name is Henry Moss and I am an Early Career Advanced Fellow in the Department of Applied Mathematics and Theoretical Physics at the University of Cambridge.
I build ML methods to tackle problems from science, working with Biologists, Chemists, Physicist, Engineers and (most recently) climate scientists. I am interested in deploying models on low-quality or low-volume experimental data. Unfortunately, I know only two tricks but they seem to work, so my research revolves around them:
- Encoding prior scientific knowledge into our ML algorithms (why bother learning it if we know it!)
- Using ML to guide the collection of a small amount of additional but high value measurements (e.g. Bayesian optimisation, active learning)
Interests
- Scalable Bayesian machine learning models to help scientists better understand the world around us
- Active learning and Bayesian optimisation to accelerate the design of new technologies
- Molecular search and gene design