evenmn

• Nipen, Thomas Nils; Haugen, Håvard Homleid; Ingstad, Magnus Sikora; Nordhagen, Even Marius; Salihi, Aram Farhad & Tedesco, Paulina Souza [Show all 14 contributors for this article] (2025). Regional data-driven weather modeling with a global stretched-grid. Artificial Intelligence for the Earth Systems (AIES). ISSN 2769-7525. p. 1–29. doi: 10.1175/aies-d-25-0001.1. Full text in Research Archive Show summary

  Abstract We present Bris, a data-driven weather model suitable for regional forecasting applications. The model extends the Artificial Intelligence Forecasting System by introducing a stretched-grid architecture that dedicates higher resolution over a regional area of interest and maintains a lower resolution elsewhere on the globe. The model is based on graph neural networks, which naturally affords arbitrary multi-resolution grid configurations. The model is applied to short-range weather prediction for the Nordics, producing forecasts at 2.5 km spatial and 6 h temporal resolution. Bris is pre-trained on 43 years of global ERA5 data at 31 km resolution and is further refined using 3.3 years of 2.5 km resolution operational analyses from the MetCoOp Ensemble Prediction System (MEPS). The performance of the model is evaluated against surface observations from measurement stations across Norway and is compared to short-range weather forecasts from MEPS. In terms of root mean squared error, Bris outperforms both the control run and the ensemble mean of MEPS for 2 m temperature. Bris also produces competitive precipitation and wind speed forecasts, but is shown to underestimate extreme events.

• Geenen, Thomas; Nordhagen, Even Marius; Sanchez, Victoria A.; O'Brien, Cathal; Lang, Simon & Alexe, Mihai [Show all 17 contributors for this article] (2025). Towards full AI model lifecycle management on EuroHPC systems, experiences with AIFS for DestinE. Procedia Computer Science. 255, p. 93–102. doi: 10.1016/j.procs.2025.02.264. Full text in Research Archive
• Camposano, Anthony Val Canillas; Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2025). Genetic Algorithm Workflow for Parameterization of a Water Model Using the Vashishta Force Field. Journal of Physical Chemistry B. ISSN 1520-6106. 129(4), p. 1331–1342. doi: 10.1021/acs.jpcb.4c06389. Full text in Research Archive Show summary

  Water participates in countless processes on Earth, and the properties of mineral surfaces can be drastically changed in the presence of water. For example, the fracture toughness of silica glass is reduced by 25% for water-filled cracks than for dry cracks [Geophys. Res.: Solid Earth 2018, 123, 9341–9354]. An accurate description of water is therefore essential for modeling the behavior of minerals in aqueous environments and, in particular, for modeling dynamic processes such as fracture, where the mechanical response of water may play an important role. On the molecular scale, molecular dynamics simulations with empirical force field methods provide a way to study large molecular systems at a relatively low computational cost. Many water models have been developed previously; however, a computationally cheap water model capable of describing reactions with minerals is lacking. Here, we present a parametrization of the water potential using the Vashishta potential form [Phys. Rev. B 1990, 41, 12197–12209]. For this 3-point water model, we obtain good agreement with experimental transport and liquid–vapor properties. Importantly, the Vashishta form opens up compatibility with existing silica glass models, thus enabling the simulation of mineral–water interactions.

• Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2025). Velocity-Dependent Dynamics of Friction and Wear. Tribology letter. ISSN 1023-8883. doi: 10.1007/s11249-025-02045-5. Full text in Research Archive
• Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2025). Tailoring Frictional Properties of Surfaces Using Diffusion Models. Journal of Physical Chemistry C. ISSN 1932-7447. 129(32), p. 14559–14564. doi: 10.1021/acs.jpcc.5c02768. Full text in Research Archive Show summary

  This work introduces an approach for precisely designing surface friction properties using a conditional generative machine learning model, specifically a denoising diffusion probabilistic model (DDPM). We created a dataset of synthetic surfaces with frictional properties determined by molecular dynamics simulations, which trained the DDPM to predict surface structures from desired frictional outcomes. Unlike traditional trial-and-error and numerical optimization methods, our approach directly yields surface designs meeting specified frictional criteria with high accuracy and efficiency. This advancement in material surface engineering demonstrates the potential of machine learning to reduce the iterative nature of surface design processes. Our findings provide a pathway for tailoring surface properties and suggest broader applications in materials science where surface characteristics are critical.

• Camposano, Anthony Val Canillas; Nordhagen, Even Marius; Malthe-Sørenssen, Anders & Sveinsson, Henrik Andersen (2025). An Extended Energy-Biased Aggregation-Volume-Bias Monte Carlo (EB-AVBMC) Method for Nucleation Simulation of a Reactive Water Potential. Journal of Chemical Theory and Computation. ISSN 1549-9618. 21(14), p. 6769–6776. doi: 10.1021/acs.jctc.5c00722. Full text in Research Archive
• Nipen, Thomas Nils; Haugen, Håvard Homleid; Ingstad, Magnus Sikora; Nordhagen, Even Marius; Salihi, Aram Farhad Shafiq & Tedesco, Paulina Souza [Show all 14 contributors for this article] (2024). Regional data-driven weather modeling with a global stretched-grid. arXiv.org. ISSN 2331-8422. p. 1–11. doi: 10.48550/arXiv.2409.02891. Full text in Research Archive
• Nordhagen, Even Marius; Kim, Jane M.; Fore, Bryce; Lovato, Alessandro & Hjorth-Jensen, Morten (2023). Efficient solutions of fermionic systems using artificial neural networks. Frontiers in Physics. 11. doi: 10.3389/fphy.2023.1061580. Full text in Research Archive Show summary

  In this study, we explore the similarities and differences between variational Monte Carlo techniques that employ conventional and artificial neural network representations of the ground-state wave function for fermionic systems. Our primary focus is on shallow neural network architectures, specifically the restricted Boltzmann machine, and we examine unsupervised learning algorithms that are appropriate for modeling complex many-body correlations. We assess the advantages and drawbacks of conventional and neural network wave functions by applying them to a range of circular quantum dot systems. Our findings, which include results for systems containing up to 90 electrons, emphasize the efficient implementation of these methods on both homogeneous and heterogeneous high-performance computing facilities.

• Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2023). Diffusion-Driven Frictional Aging in Silicon Carbide. Tribology letter. ISSN 1023-8883. 71(3), p. 1–11. doi: 10.1007/s11249-023-01762-z. Full text in Research Archive Show summary

  Abstract Friction is the force resisting relative motion of objects. The force depends on material properties, loading conditions and external factors such as temperature and humidity, but also contact aging has been identified as a primary factor. Several aging mechanisms have been proposed, including increased “contact quantity” due to plastic or elastic creep and enhanced “contact quality” due to formation of strong interfacial bonds. However, comparatively less attention has been given to other mechanisms that enhance the “contact quantity”. In this study, we explore the influence of crystal faceting on the augmentation of “contact quantity” in cubic silicon carbide, driven by the minimization of surface free energy. Our observations reveal that the temporal evolution of the frictional aging effect follows a logarithmic pattern, akin to several other aging mechanisms. However, this particular mechanism is driven by internal capillary forces instead of the normal force typically associated with friction. Due to this fundamental distinction, existing frictional aging models fail to comprehensively explain the observed behavior. In light of these findings, we derive a model for the evolution of contact area caused by diffusion-driven frictional aging, drawing upon principles from statistical mechanics. Upon application of a normal force, the friction force is increased due to plastic creep. This investigation presents an alternative explanation for the logarithmic aging behavior observed and offers the potential to contribute to the development of more accurate friction models. Graphical Abstract

View all works in NVA