High-dimensional visualisation for particle physics
by
Aula Caldirola
Dipartimento di Fisica, Università degli Studi di Milano
Physical models are typically high dimensional, while standard visualisations are restricted to be at most 3-d. In this talk I will introduce the Grand Tour, an algorithm for the visualisation of high dimensional distributions by means of dynamical low-dimensional projections. By viewing an animation of smoothly interpolated projections we can extrapolate the shape of the parameter space in high dimensions. I will then show an application to data from PDFsense, aiming to visualise the sensitivity of PDF fits to a given experiment, where the Tour reveals features not observed in other visualisations. Finally I will discuss the idea of projection pursuit, i.e. searching the high dimensional space for “interesting” low dimensional projections.