CPT

Calendar of Physics Talks Vienna

Relativistic kinetic theory with applications in astrophysics
Speaker:Paola Rioseco (ESI / Center of Mathematical Modeling U. Chile)
Abstract:I will review relativistic kinetic theory and apply it to two phenomena. The first phenomenon is accretion of matter around the Schwarzschild black hole, where we calculated its accretion rate. The second phenomenon concerns the dynamics of the kinetic gas for a thin disk in the equatorial plane of the Kerr black hole. In this case, the so-called mixing phenomenon appears, which causes that the gas configuration relaxes to stationary axisymmetric state. The talk is based on the published papers arXiv:1807.10794 / arXiv:1611.02389.
Date: Tue, 09.11.2021
Time: 14:00
Duration: 60 min
Location:Zoom: https://tuwien.zoom.us/j/97641048733?pwd=Q2JqOThkeXpUQy9ESit6R1NCZElkdz09
Contact:Romain Ruzziconi, Daniel Grumiller, Benjamin Koch

Verifying biomolecular nanoclustering in single molecule localization microscopy
Speaker:Magdalena Schneider (TU Wien, Institut für Angewandte Physik)
Abstract: Single molecule localization microscopy (SMLM) circumvents the diffraction limit of light by separating the signals from individual fluorophores in time. While the main focus of SMLM analysis up to now has been the precise localization of the fluorophores, the natural next step concerns qualitative and quantitative interpretation of the localization maps and their biological relevance. Such an interpretation is challenging, as fluorophore blinking leads to repeated detections of the same molecule. The resulting apparent localization clusters can be easily mistaken for biomolecular nanoclustering. We developed two different methods for distinguishing true biomolecular clustering from overcounting artifacts. First, a comprehensive characterization of fluorophore blinking behavior combined with Monte Carlo simulations allows for robust evaluation of localization maps with respect to true m
Date: Tue, 09.11.2021
Time: 16:00
Location:TU Wien, Institut für Angewandte Physik, E134 1040 Wien, Wiedner Hauptstraße 8-10 Yellow Tower „B“, 5th floor, SEM.R. DB gelb 05 B
Contact:Univ.Prof. Dr. Gerhard Schütz

Collinear factorisation for e+ e- collisions
Speaker: Stefano Frixione (INFN, Genoa)
Abstract:Although the computation of a vast class of e+ e- cross sections is fully doable in perturbative QED, the perturbative series is not well behaved unless a mechanism is devised that allows one to take into account, to all orders in the coupling constant, the effects due to multiple emissions of photons and e+ e- pairs off the incoming particles. I shall discuss one such mechanism, the collinear factorisation approach, that has a strict analogy with its counterpart in QCD and leads to the definition of "Parton" Distribution Functions (PDFs) of the electron and the positron. I shall present recent results that have improved the accuracy to which these PDFs are known, and the work in progress on the implementation of e+ e- collinear-factorisation formulae into the automated MadGraph5_aMC@NLO framework.
Date: Tue, 09.11.2021
Time: 16:15
Duration: 60 min
Location:ZOOM: https://univienna.zoom.us/j/93427906843?pwd=YjhSejdUVW16QjVQYUh5TVNSNFNhQT09
Contact:A. Hoang, M. Procura

Asymptotic characterization of Kerr-de Sitter and related spacetimes in all dimension
Speaker:Carlos Peón-Nieto (Salamanca)
Abstract:We review the asymptotic initial value problem in all dimensions for the Λ>0-vacuum Einstein metrics in the Fefferman-Graham picture. We give a geometric definition of the initial data in the conformally flat $\mathscr{I}$ case and a Killing initial data equation for the analytic data case.We use these results to characterize the Kerr-de Sitter family of metrics in all dimensions in terms of asymptotic initial data,which are given by a conformally flat boundary metric γ and a TT tensor canonically constructed from a particular conformal Killing vector(CKV) ξK dS of γ. The data naturally generalize for an arbitrary CKV ξ, extending the definition of the Kerr-de Sitter-like class(with conformally flat $\mathscr{I}$)to arbitrary dimensions by means of asymptotic data.We prove that each metric in the class corresponds to a(conformal)class of CKVs,i.e.the set generated from ξ by local . . .
Date: Thu, 11.11.2021
Time: 14:00
Duration: 60 min
Location: Seminarraum A, Währinger Straße 17, 2. Stock, AND via ZOOM https://univienna.zoom.us/j/6540036841?pwd=SytyVkZJZzNyRG9lMm13ejlHeHRRUT09
Contact:Piotr Chrusciel, David Fajman

Scaling Up Machine Learning For Quantum Field Theory with Equivariant Continuous Flows
Speaker:Pim de Haan (Qualcomm AI Research, Netherlands)
Abstract:Sampling from the high-dimensional probability distributions of Quantum Field Theories on a lattice using a traditional proposal distribution in Markov Chain Monte Carlo (MCMC) may suffer from critical slowing down, in which it takes many steps to generate uncorrelated samples. An interesting alternative is to learn a machine learning model to generate independent proposal samples. As many quantum field theories exhibit symmetries, it is beneficial to construct the machine learning models such that the proposal distribution is invariant to the symmetries. For the phi^4 theory, we build a neural ordinary differential equation that continuously deforms a prior distribution to the model’s proposal distribution using an equivariant vector field. Compared to previous neural methods, we greatly improve key metrics due to the design of the network and incorporation of the symmetries.
Date: Thu, 11.11.2021
Time: 16:30
Duration: 60 min
Location: https://tuwien.zoom.us/j/95482849369?pwd=SkU2WU5hMVJBQWsyVlFXRlRvQlIvdz09
Contact:Laura Donnay, Andreas Ipp

Unscrambling Entanglement through a Complex Medium
Speaker:Mehul Malik (Heriot-Watt University, Edinburgh, Scotland)
Abstract:Quantum states of light entangled in their spatial and temporal structure offer the potential for noise-robust, device-independent quantum communication networks that harness the full information carrying capacity of a photon. A central challenge in the realisation of such networks is the ability to efficiently measure and reliably transport high-dimensional entangled states of light. I will begin this talk by introducing a theoretical model for the two-photon wavefunction entangled in the position-momentum degrees-of-freedom [1]. I will go on to discuss how knowledge of this continuous wavefunction allows us to tailor discrete high-dimensional entangled states of light with a high quality in a variety of spatial mode bases. As an example, I will present the generation and measurement of high-dimensional “pixel” entanglement with fidelities exceeding 94% and entanglement dimensionalities
Date: Fri, 12.11.2021
Time: 10:00
Duration: 45 min
Location:on-line
Contact:Philipp Haslinger