A physically consistent AI-based SPH emulator for computational fluid dynamics

Amato, Eleonora; Zago, Vito; Del Negro, Ciro

doi:10.1515/nleng-2022-0359

Cited by 3 publications

(3 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The goal here is to construct a model that inherently meets specific constraints (such as periodic boundary conditions) or processes the data to mirror the real system's behavior (for example, pairwise interactions). This strategy resembles the "minimalistic" method (Alexiadis, 2023;Amato et al, 2024). However, the minimalistic method employs the network only for designated aspects of physics, such as modeling the pair potential from data.…”

Section: Physics-informed Machine Learningmentioning

confidence: 99%

“…In principle, expecting a model to perform accurately beyond its training data seems too optimistic. However, while this criterion may not generally be met, there are scenarios where, to a certain degree, this has been achieved (Alexiadis, 2023;Amato et al, 2024). Finally, we did not compare the training times of the models in this study, primarily because such comparisons are challenging to conduct on equal footing.…”

Section: Criterion 11: the Model Should Reproduce Conditions Besides ...mentioning

confidence: 99%

See 1 more Smart Citation

Designing a set of criteria for evaluating artificial neural networks trained with physics-based data to replicate molecular dynamics and other particle method trajectories

Alexiadis

2024

Front. Nanotechnol.

View full text Add to dashboard Cite

This article presents an in-depth analysis and evaluation of artificial neural networks (ANNs) when applied to replicate trajectories in molecular dynamics (MD) simulations or other particle methods. This study focuses on several architectures—feedforward neural networks (FNNs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), time convolutions (TCs), self-attention (SA), graph neural networks (GNNs), neural ordinary differential equation (ODENets), and an example of physics-informed machine learning (PIML) model—assessing their effectiveness and limitations in understanding and replicating the underlying physics of particle systems. Through this analysis, this paper introduces a comprehensive set of criteria designed to evaluate the capability of ANNs in this context. These criteria include the minimization of losses, the permutability of particle indices, the ability to predict trajectories recursively, the conservation of particles, the model’s handling of boundary conditions, and its scalability. Each network type is systematically examined to determine its strengths and weaknesses in adhering to these criteria. While, predictably, none of the networks fully meets all criteria, this study extends beyond the simple conclusion that only by integrating physics-based models into ANNs is it possible to fully replicate complex particle trajectories. Instead, it probes and delineates the extent to which various neural networks can “understand” and interpret aspects of the underlying physics, with each criterion targeting a distinct aspect of this understanding.

show abstract

Section: Physics-informed Machine Learningmentioning

confidence: 99%

Section: Criterion 11: the Model Should Reproduce Conditions Besides ...mentioning

confidence: 99%

Designing a set of criteria for evaluating artificial neural networks trained with physics-based data to replicate molecular dynamics and other particle method trajectories

Alexiadis

2024

Front. Nanotechnol.

View full text Add to dashboard Cite

show abstract

“…Nowadays, satellite remote sensing is widely employed for monitoring volcanic thermal activity globally [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21]. Numerous volcanic hotspot monitoring satellite platforms have been developed for the near real-time monitoring of thermal anomalies, such as MODVOLC [22], HOTVOLC [23], FIRMS [24], MIROVA [25] and LAV@HAZARD [26].…”

Section: Introductionmentioning

confidence: 99%

Quantitative Assessment of Volcanic Thermal Activity from Space Using an Isolation Forest Machine Learning Algorithm

Corradino,

Malaguti,

Ramsey

et al. 2024

Remote Sensing

Self Cite

View full text Add to dashboard Cite

Understanding the dynamics of volcanic activity is crucial for volcano observatories in their efforts to forecast volcanic hazards. Satellite imager data hold promise in offering crucial insights into the thermal behavior of active volcanoes worldwide, facilitating the assessment of volcanic activity levels and identifying significant changes during periods of volcano unrest. The Moderate Resolution Imaging Spectroradiometer (MODIS) sensor, aboard NASA’s Terra and Aqua satellites, provides invaluable data with high temporal and spectral resolution, enabling comprehensive thermal monitoring of eruptive activity. The accuracy of volcanic activity characterization depends on the quality of models used to relate the relationship between volcanic phenomena and target variables such as temperature. Under these circumstances, machine learning (ML) techniques such as decision trees can be employed to develop reliable models without necessarily offering any particular or explicit insights. Here, we present a ML approach for quantifying volcanic thermal activity levels in near real time using thermal infrared satellite data. We develop an unsupervised Isolation Forest machine learning algorithm, fully implemented in Google Colab using Google Earth Engine (GEE) which utilizes MODIS Land Surface Temperature (LST) data to automatically retrieve information on the thermal state of volcanoes. We evaluate the algorithm on various volcanoes worldwide characterized by different levels of volcanic activity.

show abstract