VSR-Seminar

Anfang
26.04.2019 11:30 Uhr
Ende
26.04.2019 12:30 Uhr
Veranstaltungsort
Jülich Supercomputing Centre, Hörsaal, Geb. 16.3, R. 222

Vortrag 1: High Definition Simulation of Packed Bed Chromatography

Referent: Dr. Eric von Lieres, IBG-1

Contents: Packed bed liquid chromatography is widely used in biotechnology for separating target molecules from specific impurities. The separation is facilitated by selective adsorption to functionalized particles in a column. Chromatography modeling is important for guiding experimental work through process design, optimization and uncertainty analysis. However, existing models assume homogeneity over column cross-sections and neglect particle size distributions. We utilize high definition/ high performance simulations for capturing the impact of these geometric features. Stationary velocity profiles in the interstitial column volume are described by incompressible Navier-Stokes equations. Transient concentration profiles of solute molecules are described by coupled advection-diffusion and diffusion-reaction equations in the interstitial volume and the porous particles, respectively. The model is simulated using the parallel multi-physics XNS solver, in collaboration with CATS at RWTH Aachen. XNS utilizes a stabilized space-time Galerkin finite element method. First simulations compare uniform and experimentally determined particle size distributions. These preliminary results already provide novel insights and promise high impact of further studies. However, several technical issues need to be addressed for accurately and efficiently solving realistically sized systems, as will be specified in the presentation.

Vortrag 2: Parameterizations of Spiking Neuronal Network Simulations for Coarse-Grain Simulation or Slow Time Evolution

Referent: Dr. Alexander Peyser, JSC

Contents: The team "Multiscale Simulation and Architectures" in the SimLab Neuroscience is currently involved with computational neuroscience projects for the simulation and analysis of spiking neuronal networks at several space-and-time scales, from the morphologically detailed to coarse-grain full-brain simulation involving both short-term plasticity and long-term evolution of connectivity, in addition to the development of methods for high-throughput optimization and search in this context. Two computing time allocations are closely associated with these projects: Parameter fitting for The Virtual Brain using Bayesian Inference'' (TVB, CJJSC35, 2.5M Ch), and Adaptive parameterization of structural plasticity models in neural network simulations'' (SP, CJJSC34, 1.2M Ch).

In the TVB project, we have been contributing to the development of an HPC version of the The Virtual Brain'' code and the application of Bayesian inference techniques to derive model parameters from large sets of empirical results.
The large ensembles involved are embarrassingly parallel, but the kernels themselves have been ported to take advantage of CUDA and many-core vectorization, while applying hyperparameter inference techniques to the ensembles. This project has also contributed to the Human Brain Project and The VirtualBrainCloud H2020 grants with core infrastructural components.

In the SP project, we have been extending work previously done on the slow evolution of structural plasticity for NEST simulations, accessing network behavior at new time scales, and contributing to testing of increasingly scalable implementations for HPC.

Learning-2-Learn techniques are used to optimize plasticity hyperparameters at scale using the JuPeX HPC framework developed in this context as a driver in the development of high-throughput optimization techniques. Specifically, we implement hyperparameter multiobjective optimization of the homeostatic rules of the structural plasticity algorithm for different sizes of spiking networks using fitness measures of firing rate, shape of the power spectrum and synchronization.

The team "Multiscale Simulation and Architectures'" in the SimLab Neuroscience is currently involved with computational neuroscience projects for the simulation and analysis of spiking neuronal networks at several space-and-time scales, from the morphologically detailed to coarse-grain full-brain simulation involving both short-term plasticity and long-term evolution of connectivity, in addition to the development of methods for high-throughput optimization and search in this context. Two computing time allocations are closely associated with these projects: ``Parameter fitting for The Virtual Brain using Bayesian Inference'' (TVB, CJJSC35, 2.5M Ch), and ``Adaptive parameterization of structural plasticity models in neural network simulations'' (SP, CJJSC34, 1.2M Ch).

In the TVB project, we have been contributing to the development of an HPC version of the ``The Virtual Brain'' code and the application of Bayesian inference techniques to derive model parameters from large sets of empirical results.
The large ensembles involved are embarrassingly parallel, but the kernels themselves have been ported to take advantage of CUDA and many-core vectorization, while applying hyperparameter inference techniques to the ensembles. This project has also contributed to the Human Brain Project and The VirtualBrainCloud H2020 grants with core infrastructural components.

In the SP project, we have been extending work previously done on the slow evolution of structural plasticity for NEST simulations, accessing network behavior at new time scales, and contributing to testing of increasingly scalable implementations for HPC.

Learning-2-Learn techniques are used to optimize plasticity hyperparameters at scale using the JuPeX HPC framework developed in this context as a driver in the development of high-throughput optimization techniques. Specifically, we implement hyperparameter multiobjective optimization of the homeostatic rules of the structural plasticity algorithm for different sizes of spiking networks using fitness measures of firing rate, shape of the power spectrum and synchronization.

Letzte Änderung: 17.11.2022