Description
Title
Machine Learning in IceCube
Abstract
For years, the IceCube Neutrino Observatory has collected and generated vast amounts data. However, this data is so complex and granular that traditional approaches to reconstruction are reaching their limits. This has made IceCube and similar experiments ideal candidates for machine learning (ML). Now, as these efforts mature, questions about what ML paradigms to employ and how to work with ML become salient. In this talk, I will present how optical neutrino telescopes are using ML and which directions are proving promising. I will draw parallels to other experiments to argue why IceCube is well-suited for a holistic approach to ML. Finally, I will discuss how the physics community might benefit from new ways of working with ML.
Biography
Andreas Søgaard is a Marie Skłodowska-Curie Fellow at the Niels Bohr Institute of the University of Copenhagen. Here, he is leading an effort to develop of common, open-source graph neural network tools for event reconstruction in IceCube and similar neutrino telescope experiments. He received his PhD in Experimental Particle Physics from the University of Edinburgh, working on the ATLAS experiment at CERN. After graduating we was a postdoctoral research associate at the University of Edinburgh before spending 2+ years in the private sector, as a Partner and Chief AI Officer in a Danish data science consultancy.
Machine Learning in IceCube
Abstract
For years, the IceCube Neutrino Observatory has collected and generated vast amounts data. However, this data is so complex and granular that traditional approaches to reconstruction are reaching their limits. This has made IceCube and similar experiments ideal candidates for machine learning (ML). Now, as these efforts mature, questions about what ML paradigms to employ and how to work with ML become salient. In this talk, I will present how optical neutrino telescopes are using ML and which directions are proving promising. I will draw parallels to other experiments to argue why IceCube is well-suited for a holistic approach to ML. Finally, I will discuss how the physics community might benefit from new ways of working with ML.
Biography
Andreas Søgaard is a Marie Skłodowska-Curie Fellow at the Niels Bohr Institute of the University of Copenhagen. Here, he is leading an effort to develop of common, open-source graph neural network tools for event reconstruction in IceCube and similar neutrino telescope experiments. He received his PhD in Experimental Particle Physics from the University of Edinburgh, working on the ATLAS experiment at CERN. After graduating we was a postdoctoral research associate at the University of Edinburgh before spending 2+ years in the private sector, as a Partner and Chief AI Officer in a Danish data science consultancy.
The A3D3 Seminar is a monthly lecture series that hosts scholars working across applied areas of artificial intelligence, such as hardware algorithm co-development, high energy physics, multi-messenger astrophysics, and neuroscience. Our presenters come from all four domain fields and include occasional external speakers beyond the A3D3 science areas, governmental agencies and industry. The seminar will be recorded and published in YouTube. To receive future event update, subscribe here.
Organised by
Matthew Graham, Kate Scholberg