Group seminar Theoretical Physics III

11.11.2020 10:45 a.m., room: S-439
Dr. Jernej Mravlje (Jožef Štefan Institute, Ljubljana, Slovenia)
Charge and spin-transport at high temperatures
18.11.2020 10:45 a.m., room: S-439
Nico Unglert, M.Sc.
The nodal-line semimetal ZrSiTe under external pressure - a DFT study
25.11.2020 10:45 a.m., room: S-439
Andreas Weh, M.Sc.
Half-metals in the presence of local many-body correlations
2.12.2020 10:45 a.m., room: S-439
Alyn James, MSc. (University of Bristol, UK)
Calculating DFT+DMFT wavefunction-dependent quantities using the new Elk-TRIQS interface
9.12.2020 10:45 a.m., room: S-439
Dr. Johannes Thiele (CEA, Saclay, France)
Neuromorphic computing: Overcoming computational challenges in AI with innovative hardware paradigms

Machine learning, and in particular a class of algorithms called Deep Learning, has revolutionized a large number of application domains in recent years. However, the increased capacities of these algorithms come at a price: ever increasing demands for computation. This poses today two fundamental problems: firstly, training of state-of-the-art models has become prohibitively expensive, with only a handful of actors worldwide being capable of training the most advanced models from scratch. Secondly, even when these models are trained, their energy and memory consumption is incompatible with the constraints of most embedded systems, which hinders their use in mobile and autonomous systems. At the same, Moore’s law, which has for a long time been a guarantor for the continuing growth of computation power, seems to be stagnating. Therefore, the rise of Deep Learning has caused a renaissance in the search for novel computing paradigms: algorithm developers, IC designers, neuroscientists and material scientists are starting to work together to surmount the limits of current digital hardware. This talk will give a high level introduction to the current advances and challenges in the field of machine learning, with a focus on the different hardware paradigms that are proposed to overcome these problems.

16.12.2020 10:45 a.m., room: S-439
Priv.-Doz. Dr. Andreas Sinner (University of Augsburg)
Pairing transition in a double layer with interlayer Coulomb repulsion
20.1.2021 10:45 a.m., room: S-439
Dr. Firat Yilmaz
Thermal Hall conductance in time-reversal symmetry broken systems: A case study, impurity mechanism in chiral superconductors
17.2.2021 10:45 a.m., room: S-439
Jernej Mravlje (IJS, Ljubljana)
The role of correlations and thermal disorder in transport properties of iron in the Earth's core
** Datum korrigiert: Mittwoch, 17. Februar 2021 **