Friday 24 May 2024 |
Events for day: Monday 06 May 2024 |
11:00 - 12:00 Journal Club: Experiment & Phenomenology – meeting Improving Parametric Neural Networks in High Energy Physics School PARTICLES AND ACCELERATORS Abstract: Signal-background classification is a central problem in High-Energy Physics (HEP), that plays a major role for the discovery of new fundamental particles. Recently machine learning algorithms and specially neural networks have become very popular for this matter. Parametric Neural Network (pNN) leverages multiple signal mass hypotheses as an additional input feature to effectively replace a whole set of individual classifiers, each providing (in principle) the best response for the corresponding mass hypothesis. In this talk we aim at deepening the understanding of pNNs in light of real-world usage. We further propos ... 14:30 - 16:00 Lecture Controlling Growth Rate of Rational Points by Systole School MATHEMATICS Let $X$ be a complex ball quotient by a nonuniform neat lattice in $PU(n,1)$. Using hyperbolic geometry, we establish a uniform lower bound on the volume of subvarieties of $X$ in terms of the systole, a geometric quantity that measures the shortest closed geodesic in $X$. We combine this result with the strategy of Bombieri-Pila to show that if the toroidal compactification of $X$ is defined over a number field $K,$ then, under a mild assumption on $X,$ the systole of $X$ controls the growth rate of $K$-rational points on $X$. Online via zoom: Meeting ID: 908 611 6889 Passcode: 362880 Venue: Niavaran, ... |