Tilman Plehn, Heidelberg University [HET Seminar], Machine Learning with Precision and Error Bars

Thursday 20 Mar 2025, 11:00 → 12:00 US/Eastern

CFNS Library (https://bnl.zoomgov.com/j/1605506805?pwd=AA9rwL4hM5AwG2b0zEwDdXjzhvJjSz.1)

LHC as one of the most data-intensive scientific endeavours provides
the perfect link between fundamental physics research and modern data
science. As machine learning is transforming our lives,
literally, no aspect of LHC physics is left untouched. This starts
with identifying data for classic or optimal analyses and extends to
anomaly searches and powerful simulations based on perturbative
quantum field theory. I will give a few examples for the
transformative power of modern machine learning in particle physics,
show how our understanding of uncertainties adds new flavors to
machine learning, and explain how generative neural networks allow us
to realize our dream of making LHC data available to a broad
scientific community.