Refine
Year of publication
Document Type
- In Proceedings (204)
- Article (150)
- Master's Thesis (41)
- ZIB-Report (37)
- Other (14)
- Book chapter (13)
- Doctoral Thesis (10)
- Bachelor's Thesis (8)
- Proceedings (4)
- Poster (3)
Keywords
- DHT (4)
- Paxos (4)
- Peer-to-Peer (3)
- Consensus (2)
- GPGPU (2)
- HR-pQCT measurement (2)
- MPI (2)
- Nebenläufigkeitskontrolle (2)
- Strukturierte Overlay-Netzwerke (2)
- Verteilte Transaktionen (2)
Institute
- Distributed Algorithms and Supercomputing (487) (remove)
The microphysical structure of the lunar regolith provides information on the geologic history of the Moon. We used remote sensing measurements of thermal emission and a thermophysical model to determine the microphysical properties of the lunar regolith. We expand upon previous investigations by developing a microphysical thermal model, which more directly simulates regolith properties, such as grain size and volume filling factor. The modeled temperatures are matched with surface temperatures measured by the Diviner Lunar Radiometer Experiment on board the Lunar Reconnaissance Orbiter. The maria and highlands are investigated separately and characterized in the model by a difference in albedo and grain density. We find similar regolith temperatures for both terrains, which can be well described by similar volume filling factor profiles and mean grain sizes obtained from returned Apollo samples. We also investigate a significantly lower thermal conductivity for highlands, which formally also gives a very good solution, but in a parameter range that is well outside the Apollo data. We then study the latitudinal dependence of regolith properties up to ±80° latitude. When assuming constant regolith properties, we find that a variation of the solar incidence-dependent albedo can reduce the initially observed latitudinal gradient between model and Diviner measurements significantly. A better match between measurements and model can be achieved by a variation in intrinsic regolith properties with a decrease in bulk density with increasing latitude. We find that a variation in grain size alone cannot explain the Diviner measurements at higher latitudes.
From August 2014 to September 2016, the ESA operated the Rosetta spacecraft mission alongside with comet 67P/Churyumov-Gerasimenko (67P). The mission provided valuable long-term data on the comet’s nucleus, including its volume, mass, tensor of inertia, spatial position of the orbital trajectory, and rotational state.
A generalized understanding of protein dynamics is an unsolved scientific problem, the solution of which is critical to the interpretation of the structure-function relationships that govern essential biological processes. Here, we approach this problem by constructing coarse-grained molecular potentials based on artificial neural networks and grounded in statistical mechanics. For training, we build a unique dataset of unbiased all-atom molecular dynamics simulations of approximately 9 ms for twelve different proteins with multiple secondary structure arrangements. The coarse-grained models are capable of accelerating the dynamics by more than three orders of magnitude while preserving the thermodynamics of the systems. Coarse-grained simulations identify relevant structural states in the ensemble with comparable energetics to the all-atom systems. Furthermore, we show that a single coarse-grained potential can integrate all twelve proteins and can capture experimental structural features of mutated proteins. These results indicate that machine learning coarse-grained potentials could provide a feasible approach to simulate and understand protein dynamics.
Regolith is formed through weathering of the local rock by meteorite bombardment, space weathering (Pieters & Noble, 2016) and thermal erosion (Delbo et al., 2014). In the case of the Moon, the space weathering effects and diurnal temperature variations are reduced towards the poles. The aim of this study is to investigate whether the lunar regolith properties derived from the comparison of regolith temperatures measured by the Diviner radiometer (Paige et al., 2010) on board the Lunar Reconnaissance Orbiter (LRO) with simulated temperatures derived from a microphysical thermal model show a latitudinal dependence. The developed microphysical thermal model expands upon previous models by more directly simulating regolith properties, such as grain radius and volume filling factor.
The successful recent application of machine learning methods to scientific problems includes the learning of flexible and accurate atomic-level force-fields for materials and biomolecules from quantum chemical data. In parallel, the machine learning of force-fields at coarser resolutions is rapidly gaining relevance as an efficient way to represent the higher-body interactions needed in coarse-grained force-fields to compensate for the omitted degrees of freedom. Coarse-grained models are important for the study of systems at time and length scales exceeding those of atomistic simulations. However, the development of transferable coarse-grained models via machine learning still presents significant challenges. Here, we discuss recent developments in this field and current efforts to address the remaining challenges.
Machine-learned coarse-grained (CG) models have the potential for simulating large molecular complexes beyond what is possible with atomistic molecular dynamics. However, training accurate CG models remains a challenge. A widely used methodology for learning bottom-up CG force fields maps forces from all-atom molecular dynamics to the CG representation and matches them with a CG force field on average. We show that there is flexibility in how to map all-atom forces to the CG representation and that the most commonly used mapping methods are statistically inefficient and potentially even incorrect in the presence of constraints in the all-atom simulation. We define an optimization statement for force mappings and demonstrate that substantially improved CG force fields can be learned from the same simulation data when using optimized force maps. The method is demonstrated on the miniproteins chignolin and tryptophan cage and published as open-source code.