Sommerfeld Theory Colloquium (ASC)

Machine learning using artificial neural networks is revolutionizing many areas of science and technology. This increases the urgency for exploring alternatives to artificial neural networks running on digital hardware. These alternatives might eventually be faster and/or more power-efficient. With this in mind, we ask the question whether one can identify a general principle that would enable a nonlinear physical system to become a self-learning machine - i.e. a physical information-processing device where internal degrees of freedom self-adjust by physical interactions to learn a desired input-output relation. In this talk, I will present our recent idea on how this might be achieved for arbitrary time-reversal-invariant Hamiltonian systems. I will introduce the principle of 'Hamiltonian Echo Backpropagation', and demonstrate how efficient learning could be possible in a wide class of physical systems.

See: Self-learning Machines based on Hamiltonian Echo Backpropagation,

Victor Lopez-Pastor, Florian Marquardt, arXiv 2103.04992 (2021)

Tensor Network States, like matrix product or projected entangled pair states play an important role in both, quantum information theory and many-body physics. They offer a compact and efficient representation, enabling accelerated numerical computations and providing intuitive insights into many-body phenomena. In this talk, I will discuss how certain states can be efficiently prepared and manipulated using quantum devices, highlighting the use of local operations and classical communication. Tensor networks can also be used to efficiently describe quantum channels. I will also mention how those channels can be efficiently implemented as quantum circuits.

The decay of a false vacuum of unbroken B-L, the difference of baryon and lepton number, is an intriguing and testable mechanism to gen- erate the initial conditions of the hot early universe. If B-L is broken at the grand unification scale, the false vacuum phase yields hybrid inflation, ending in tachyonic preheating. The dynamics of the B-L breaking Higgs field and thermal processes produce an abundance of heavy neutrinos whose decays generate entropy, baryon asymmetry and dark matter.

After a short introduction to the swampland program and the challenges one faces in explicit tests of some conjectures, this talk will focus on the computation of periods and their application to the swampland program. A detailed understanding of periods allows to answer questions about the existence of solutions or the finiteness of the string landscape, as well as explicit constructions of models. Moreover, applications of periods outside of string compactifications are discussed, such as the computation of Feynman integrals, the tameness of QFTs and the appearance of periods in many other physical systems.

The knowledge of all correlation functions of a system is equivalent to solving the corresponding quantum many-body

problem. If one can identify the relevant degrees of freedom, the knowledge of a finite set of correlation functions is in

many cases sufficient to determine a sufficiently accurate solution of the corresponding field theory. Complete

factorization is equivalent to identifying the relevant degrees of freedom where the Hamiltonian becomes diagonal. I

will give examples how one can apply this powerful theoretical concept in experiment.

A detailed study of non-translation invariant correlation functions reveals that the pre-thermalized state a system of

two 1-dimensional quantum gas relaxes to after a splitting quench [1], is described by a generalized Gibbs ensemble

[2]. This is verified through phase correlations up to 10th order.

Interference in a pair of tunnel-coupled one-dimensional atomic super-fluids, which realize the quantum Sine-Gordon /

massive Thirring models, allows us to study if, and under which conditions the higher correlation functions factorize

[3]. This allowed us to characterize the essential features of the model solely from our experimental measurements:

detecting the relevant quasi-particles, their interactions and the different topologically distinct vacuum-states the

quasi-particles live in. The experiment thus provides a comprehensive insight into the components needed to solve a

non-trivial quantum field theory.

Our examples establish a general method to analyse quantum systems through experiments. It thus represents a

crucial ingredient towards the implementation and verification of quantum simulators.

Work performed in collaboration with E.Demler (Harvard), Th. Gasenzer und J. Berges (Heidelberg).

Supported by the Wittgenstein Prize, the Austrian Science Foundation (FWF): SFB FoQuS: F40-P10 and

the EU: ERC-AdG QuantumRelax

[1] M. Gring et al., Science, 337, 1318 (2012);

[2] T. Langen et al., Science 348 207-211 (2015).

[3] T. Schweigler et al., Nature 545, 323 (2017), arXiv:1505.03126

In February 2016 the LIGO team announced the detection of gravitational

waves (GW) created by the merger of two black holes. In addition

to confirming a major prediction of general relativity, successful

GW detection would provide a powerful new tool for astrophysics.

Given their evident importance, the LIGO results and the methods

which led to them deserve independent critical analysis. This talk

will present the results of one such study in a manner suitable for

non-specialists.

I will describe the process of how to award Nobel Prizes and then give the history behind this year’s prize.

Understanding dark energy and black holes remain a great challenge to fundamental physics. In this talk I will review the difficulties and explore some new and speculative approaches.

I will discuss Effective Field Theories that can originate from microscopic unitary theories, and their relation to moment theory. I will show that massive gravity, theories with isolated massive higher-spin particles, and theories with very irrelevant interactions, don't posses healthy UV completions, and I will show how Vector Meson Dominance follows from such first principles.

With Run II, the LHC experiment already reached an unprecedented

level of precision compared to previous hadron colliders. The amount

of data collected in Run III and the High-Luminosity run will increase

quickly and dramatically. I will discuss specific challenges to theorists

that must be overcome to provide predictions that have the precision

necessary to match the accuracy of data. A few novel ideas to exploit

this precision directly or indirectly will also be discussed.