fizinfo AT lists.kfki.hu
Subject: ELFT HÍRADÓ
List archive
- From: Nagy Attila <sem-admin AT szfki.hu>
- To: SZFI User <szfkiuser AT szfki.hu>, Fizinfo <fizinfo AT lists.kfki.hu>
- Subject: [Fizinfo] Szemináriumok - Seminars: Dominik Sulz
- Date: Mon, 27 May 2024 07:29:49 +0200
- Authentication-results: smtp012.wigner.hu (amavisd-new); dkim=pass (1024-bit key) reason="pass (just generated, assumed good)" header.d=szfki.hu
SZFI Seminar
DOMINIK SULZ
University of Tübingen (host: Legeza Örs)
_MANY-BODY QUANTUM DYNAMICS USING TREE TENSOR NETWORKS_
Tuesday, 28 May 2024, 13:00, KFKI Campus, Bldg. 1, 2nd floor, Conference Room
In many applications, it is of interest to numerically solve high-order tensor differential equations, e.g. the many-body Schrödinger equation in quantum physics or discretized high-dimensional time-dependent partial differential equations. Tree tensor networks (TTNs) turned out to be a promising ansatz to overcome the so-called "curse of dimensionality'' as they are an efficient hierarchical data sparse format to approximate tensors. Based on the dynamical low-rank approximation, several novel time integration methods for TTNs have been proposed in the last few years. In a recursion from the leaves up to the root, all nodes are updated by only solving small matrix-/tensor differential equations. The proposed methods are robust even in the presence of small singular values. Further properties like rank-adaptivity, norm and energy preservation and error behaviour are mathematically are discussed.
The talk intends to give an overview of the latest results for dynamical low-rank approximations. Further, the different time integration methods and their applicability to many-body quantum systems are discussed. Numerical experiments validate the theoretical results and show several applications from quantum physics.
MINDEN ÉRDEKLŐDŐT SZÍVESEN LÁTUNK! - EVERYONE IS WELCOME TO ATTEND.
Attila Nagy
sem-admin AT szfki.hu
- [Fizinfo] Szemináriumok - Seminars: Dominik Sulz, Nagy Attila, 05/27/2024
Archive powered by MHonArc 2.6.19+.