Choose timezone
Your profile timezone:
Description
TITLE: Searching for new physics without knowing the signal model
Despite the overwhelming success of current fundamental theories, these cannot, for example, describe quantum gravity, account for dark matter, or explain the absence of antimatter in our universe. In exploring such fundamental puzzles, (astro)+particle physics implements Machine Learning across a very wide range of research directions.
Our work in recent years deals with the following questions:
How can we provide faster, precise simulations of statistical and quantum processes? How do we find anomalies in vast and complex data sets? How do we optimally build and test models for new (astroparticle) physics? How do we construct hardware-aware, efficient data-analysis pipelines? Can we combine differentiable programming and optimisation strategies to improve physical algorithms?
I will discuss some recent examples from our work and suggestions for the future.
