TTK Seminars WS 22/23

Thu 13.10.2022, 16.30 h

Guilhem Lavaux (Institut d’Astrophysique de Paris)

Fundamental physics without cosmic variance: exploring the universe through the BORG framework and the Sibelius Dark simulation

Cosmic variance poses limits to the analyses of the physics of the nearby Universe. For example, it hinders establishing an accurate observational link between halo formation history and galaxy properties or constraining the physics of Dark Matter. The BORG framework offers a way out by generating constrained initial conditions which reproduce all the Large scale structures in a simulated environment with minimal effects due to cosmic variance. By relying on the physics of gravitational interaction and our knowledge of the Universe on large scales, it can generate an ensemble of plausible samples of the initial conditions of our Universe. That framework unlocks the possibility of studying different models of galaxy formation and their impact on different observed environments. The physical environment is notably responsible for the specific mass accretion history of small-scale structures. The Sibelius-Dark simulation is an effort in that direction. I will showcase some direct results obtained with this simulation, and also more generally by the BORG framework on e.g. gravitational lensing and upper bound on the cross-section of Dark Matter particles.

Host: J. Lesgourgues


Thu 20.10.2022, 16.30 h (via zoom)

Saša Prelovšek (University of Ljubljana)

Exotic and conventional hadrons from lattice QCD

I will review how strongly stable and strongly decaying hadrons are extracted from lattice QCD. Few recent results on exotic and conventional hadrons will be reviewed. Among others, I'll discuss the first lattice results on the doubly charm tetraquark, which was discovered in 2021 and is the longest lived exotic hadron so far.

Host: R. Harlander


Thu 27.10.2022, 16.30 h

Azadeh Moradinezhad (University of Geneva)

Fundamental Physics with galaxy clustering beyond the power spectrum

The cosmic large-scale structure (LSS) provides a rich trove of information to probe the origin of the universe, its evolution and its constituents. The upcoming wide-field galaxy surveys will provide unprecedented volume of high precision data, promising to constrain deviation from LCDM model at percent level to search for new physics. Given the nonlinear nature of the LSS, extracting the non-Gaussian information not fully captured by the galaxy power spectrum is key in fully realizing the potential of upcoming galaxy surveys. Furthermore, most types of primordial non-Gaussianity do not affect the observed galaxy power spectrum and can only be probed with clustering statistics beyond the power spectrum. In this talk, I will describe several summary statistics (including galaxy bispectrum, marked power spectrum, weighted skew spectra, and wavelet moments) that capture this non-Gaussian information. I will illusterate their cosmological information content, as well as the associated challenges in their measurments and interpretation. Finally, I will discuss prospects of applying these statistics to upcoming data to constrain LCDM model, neutrino mass, modification to gravity and primordial non-Gaussianity.

Host: J. Lesgourgues


Thu 17.11.2022, 16.30 h

Enrico Peretti (NBI Copenhagen)

Particle acceleration and multimessenger radiation from wind bubbles

Young massive stellar clusters as well as starburst galaxies and active galactic nuclei can indeed launch and sustain powerful outflows featuring very high velocity and large opening angle. Such winds often develop a bubble structure characterized by an inner wind shock and an outer forward shock. During the time the forward shock expands in the surrounding medium, the inner wind shock quickly decelerates while remaining strong, thereby creating ideal conditions for particle acceleration.

I will discuss diffusive shock acceleration as key process to energize particles in wind bubbles and I will highlight the associated multimessenger implications in terms of high energy photons, neutrinos and escaping cosmic rays.

Host: P. Mertsch


Thu 24.11.2022, 16.30 h

Giada Peron (MPI für Kernphysik Heidelberg)

Probing Cosmic ray density with giant molecular clouds

The accumulation and effective mixture of relativistic particles during their propagation through the interstellar magnetic fields results in the formation of the so-called "sea" of galactic Cosmic Rays (CRs).  The level and the energy distribution of the CR sea is determined by the operation of all galactic accelerators over the confinement time of CRs in the Galactic Disk. The homogeneity of CRs, however, can be violated on smaller scales, in the form of excess fluxes over the CR sea, caused by the injection of fresh relativistic particles by young accelerators into the interstellar medium (ISM). CRs interacting with Giant Molecular Clouds with masses larger than 105 Solar masses produce “enhanced” gamma-ray emission which can be used to probe the level of the CR sea throughout the Galactic Disk. I present the results that we obtained from Fermi-LAT analysis of molecular clouds located at different distances from the Galactic Center. I will also discuss their connection with lower energy tracers of CRs and the prospect for detection of clouds and diffuse emission at high and very high energies.

Host: P. Mertsch


Thu 15.12.2022, 16.30 h

Diego Blas (IFAE and Universitat Autonoma de Barcelona)


Host: J. Lesgourgues


Thu 19.01.2023, 16.30 h

Claudius Krause (Uni Heidelberg)


Host: T. Finke


Thu 26.01.2023, 16.30 h

Daniel Figueroa (University of Valencia)


Host: J. Lesgourgues


Thu 02.02.2023, 16.30 h

Thea Aarrestad (ETH Zürich)

Ultrafast Machine Learning Inference at the Large Hadron Collider

At the CERN Large Hadron Collider, protons are brought to collide hundreds of millions of times per second. The collision debris allows us to study the fundamental building blocks of the universe and look for hints of new forces and particles. The vast majority of the collision data are immediately discarded by a real-time event filtering system due to storage and computational limitations. While most of these data are uninterresting, signals of new physics might be inadvertendly thrown away in the process. The first stage of this event filtering system consists of hundreds of field programmable gate arrays (FPGAs), tasked with rejecting over 98% of the proton collisions within a few microseconds. With the start of High Luminosity LHC in 2029, a more granular detector and more particles per collision will increase the event complexity significantly, and ultimately require the FPGA farm to process an amount of data comparable to 5% of the total internet traffic.

In this talk, I will discuss how real-time Machine Learning (ML) is used to process and filter this enormous amount of data in order to improve physics acceptance, and how ML can be used to select data in ways never before performed at colliders.