Sommerfeld Theory Colloquium (ASC)

Learning algorithms using deep neural networks are currently having a major impact on basic sciences. The physics of complex quantum systems is no exception, with multiple applications that constitute a new field of research. Examples include the representation and optimization of wave functions of quantum systems with large numbers of degrees of freedom (neural quantum states), the determination of wave functions from measurements (quantum tomography), and applications to the electronic structure of materials, such as the determination of more precise density functionals or the learning of force fields to accelerate molecular dynamics simulations. I will survey some of these applications, with an emphasis on neural quantum states.

I will discuss recent progress in the study of cosmological applications of string compactifications with stabilised moduli, focusing in particular on inflation, reheating and dark energy.

Since the discovery of the first binary black-hole merger in 2015, analytical and numerical solutions to the relativistic two-body problem have been essential for the detection and interpretation of more than 100 gravitational-wave signals from compact-object binaries. Future experiments will detect black holes at cosmic dawn, probe the nature of gravity and reveal the composition of neutron stars with exquisite precision. Theoretical advances (of up to two orders of magnitude in the precision with which we can predict relativistic dynamics) are needed to turn gravitational-wave astronomy into precision laboratories of astrophysics, cosmology, and gravity. In this talk, I will discuss recent advances in modeling the two-body dynamics and gravitational radiation, review the science that accurate waveform models have enabled with LIGO-Virgo gravitational-wave observations, and highlight the theoretical challenges that lie ahead to fully exploit the discovery potential of increasingly sensitive detectors on the ground, such as the Einstein Telescope and Cosmic Explorer, and in space, such as the Laser Interferometer Space Antenna (LISA).

Throughout the century that has passed since Ernst Ising submitted his PhD thesis in 1924, the Ising (-Lenz) model has provided an incredibly fruitful challenge that gave rise to entirely new branches of physics and mathematics. In this colloquium I will focus on the conformal bootstrap program which was designed by Polyakov in 1974 as a mathematical method to access non-perturbative aspects of critical systems/fixed points of the renormalization groups. In the light of holography, such systems are also relevant for the study of quantum gravity. In my presentation I will review some of the milestone achievements of the modern conformal bootstrap and outline current frontiers. The advances will be benchmarked mostly within the context of the 3D Ising model.

Perturbation theory remains one of the main tools in physics, in particular in quantum theories. However, most perturbative series diverge factorially, and it is not obvious how to extract information from them. Their divergence also suggests that, in order to obtain accurate results, one might need additional non-perturbative information. The theory of resurgence has been proposed as a general framework to address these issues. In this talk I will give an introduction to this theory and will illustrate it with applications -old and new- in quantum mechanics, quantum field theory and string theory.

String theory is around 50 years old and for much of that time it has been proclaimed as a quantum theory of gravity unified with all forces and matter. However, we still don’t know its fundamental formulation, although we do now know it is not just a theory of strings. Nonetheless, it has led to many new and surprising insights, with concepts that were once seen as absolute now seen as dependent on the “duality frame". In this talk I survey some of these insights and discuss their implications for physics and the fundamental formulation of string theory.

Strongly correlated electron systems, i.e. systems where the interaction between electrons cannot be treated as an effective potential, are an extremely fascinating, but also very challenging topic in modern solid state physics. The challenge arises in parts due to the simultaneous importance of non-local kinetic and local correlation effects, which make it important to treat both at equal footing. For this reason Dynamical Mean Field Theory (DMFT) has in the last decades become the state of the art method for electronic structure caluclations of strongly correlated electrons as it includes local correlation effects exactly, but also respects kinetic effects in terms of an embedding approach. In this talk, we will first motivate our interest in strongly correlated materials by giving an example regarding the fascinating properties that these materials can exhibit. This will be followed by an intuitive introduction to DMFT. Finally we present results from our DMFT studies of two strongly correlated systems: BaOsO3 [1] and tetragonal CuO (t-CuO) [2]. [1] MB, Jernej Mravlje, Martin Grundner, Ulrich Schollwoeck, and Manuel Zingl, Phys. Rev. B. 103, 165133 (2021) [2] MB, B. Bacq-Labreuil, M. Grundner, S. Biermann, U. Schollwoeck, S. Paeckel, and B. Lenz, SciPost Phys., 14, 010 (2023)

Electron-positron pair production in ultra-strong electric fields, the Sauter-Schwinger effect, is a long-standing theoretical prediction. In this talk the Sauter-Schwinger effect will be introduced and the related field-strength and energy scales as well as the possibility to verify this effect in upcoming multi-petawatt laser facilities will be discussed. The Dirac-Heisenberg-Wigner formalism provides a fully Poincaré-covariant, non-perturbative phase space description of the Sauter-Schwinger effect, and therefore its key quantities will be introduced. Some respective numerical results will be shown and discussed.

An interpretation of a particle distribution at finite (non-asymptotic) times will be provided via a Gedankenexperiment. This in turn enables one to isolate and, therefore, identify the relevant time scales of particle formation. The resulting generic aspects for particle creation in quantum physics beyond perturbation theory will be elucidated.

The life of a protein, from birth till death, is complex and challenging. At times, because of stresses or bad luck, it might take the wrong conformation and start aggregating. This process is intrinsic to the physics of proteins, and life has had to cope with it since its early days. The solution devised by evolution comes in the form of chaperone protein, a broad class of machines, present in all organisms on Earth, that repair conformationally damaged proteins, making them functional again, at an energy cost. In this talk I will provide a view of our present understanding of the molecular mechanism of function of Hsp70, possibly the most central of all chaperones, and of its consequences on proteins.

We study universal traits which emerge both in real-world complex datasets, as well as in artificially generated ones. Our approach is to analogize data to a physical system and employ tools from statistical physics and Random Matrix Theory (RMT) to reveal their underlying structure. We focus on the feature-feature covariance matrix, analyzing both its local and global eigenvalue statistics. Our main observations are: (i) The power-law scalings that the bulk of its eigenvalues exhibit are vastly different for uncorrelated random data compared to real-world data, (ii) this scaling behavior can be completely recovered by introducing long range correlations in a simple way to the synthetic data, (iii) both generated and real-world datasets lie in the same universality class from the RMT perspective, as chaotic rather than integrable systems, (iv) the expected RMT statistical behavior already manifests for empirical covariance matrices at dataset sizes significantly smaller than those conventionally used for real-world training, and can be related to the number of samples required to approximate the population power-law scaling behavior, (v) the Shannon entropy is correlated with local RMT structure and eigenvalues scaling, and substantially smaller in strongly correlated datasets compared to uncorrelated synthetic data, and requires fewer samples to reach the distribution entropy. These findings can have numerous implications to the characterization of the complexity of data sets, including differentiating synthetically generated from natural data, quantifying noise, developing better data pruning methods and classifying effective learning models utilizing these scaling laws.