Regensburg 2019 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

TT: Fachverband Tiefe Temperaturen

TT 44: Correlated Electrons: Method Development

TT 44.6: Vortrag

Mittwoch, 3. April 2019, 16:15–16:30, H7

Spektra: An Online Tool for Analytic Continuation — •Khaldoon Ghanem — Max Max Planck Institute for Solid State Research, Stuttgart, Germany

The stochastic sampling method (StochS) is used for the analytic continuation of quantum Monte Carlo data from the imaginary axis to the real one. Due to its lack of explicit parameters and its potential of resolving sharp features, StochS provides a promising alternative to the commonly used maximum entropy method.

In earlier DPG talks, we presented an efficient algorithm of stochastic sampling that reduces the computational cost by orders of magnitude. We also showed that StochS has an implicit default model given by the discretization grid and provided a recipe for choosing this grid. We extended StochS into a gridless method (gStochS) by sampling the grid points from a default model and then extend gStochS to sample over a whole class of default models with different widths (eStochS).

In this talk, we review the aforementioned developments in a unifying framework where stochastic sampling methods are organized as a hierarchy of increasing power and cost. We present a highly-optimized implementation of all three methods: StochS, gStochS and eStochS, and make it accessible through an online interface. The user can simply upload data in a convenient format and submit calculations with few clicks. The computation is done remotely and results are available in seconds/minutes. The interface also facilitates comparing different results in order to apply our recipe for choosing the optimal hyper-parameters, if needed.