SEARCH · Engineering Papers

Results for “factorization machine”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 649 records · Page 36

All sources NASA NTRS DOE OSTI

DOE OSTI2014

s-Step Krylov Subspace Methods as Bottom Solvers for Geometric Multigrid

Geometric multigrid solvers within adaptive mesh refinement (AMR) applications often reach a point where further coarsening of the grid becomes impractical as individual sub domain sizes approach unity. At this point the most common solution is to use a bottom solver, such as BiCGStab, to reduce the residual by a fixed factor at the coarsest level. Each iteration of BiCGStab requires multiple global reductions (MPI collectives). As the number of BiCGStab iterations required for convergence grows with problem size, and the time for each collective operation increases with machine scale, bottom solves in large-scale applications can constitute a significant fraction of the overall multigrid solve time. In this paper, we implement, evaluate, and optimize a communication-avoiding s-step formulation of BiCGStab (CABiCGStab for short) as a high-performance, distributed-memory bottom solver for geometric multigrid solvers. This is the first time s-step Krylov subspace methods have been leveraged to improve multigrid bottom solver performance. We use a synthetic benchmark for detailed analysis and integrate the best implementation into BoxLib in order to evaluate the benefit of a s-step Krylov subspace method on the multigrid solves found in the applications LMC and Nyx on up to 32,768 cores on the Cray XE6 at NERSC. Overall, we see bottom solver improvements of up to 4.2x on synthetic problems and up to 2.7x in real applications. This results in as much as a 1.5x improvement in solver performance in real applications.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Position Papers for Inverse Methods for Complex Systems under Uncertainty Workshop

The ability to solve inverse problems – inferring unknown parameters, structures, or states of a system from observed data – is essential for advancing scientific discovery and innovation capabilities for the DOE mission. Basic research needs and challenges are particularly acute in emerging areas such as the interactive, data-driven, modeling and simulation of digital twins; decision support for experiments at DOE scientific user facilities; and for other complex systems and workflows. Inverse problems are at the heart of understanding and controlling complex systems due to factors such as observational data with varying modalities and fidelities, inherent uncertainties in physical measurements and numerical models, and the computational demands of rapid and high-fidelity simulations. The convergence of recent scientific computing trends – scientific machine learning, artificial intelligence, and computing advances such as exascale computing – is creating unprecedented opportunities. These advancements offer the potential to revolutionize how we approach inverse problems to extract actionable insights with the required level of accuracy and computational efficiency. This workshop and the Call for Position Papers are vital steps in bringing together experts to collectively explore and identify the new computational and mathematical directions needed in inverse methods for complex systems under uncertainty.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2016

SU-F-T-485: Independent Remote Audits for TG51 NonCompliant Photon Beams Performed by the IROC Houston QA Center

Purpose: IROC-H conducts external audits for output check verification of photon and electron beams. Many of these beams can meet the geometric requirements of the TG 51 calibration protocol. For those photon beams that are non TG 51 compliant like Elekta GammaKnife, Accuray CyberKnife and TomoTherapy, IROC-H has specific audit tools to monitor the reference calibration. Methods: IROC-H used its TLD and OSLD remote monitoring systems to verify the output of machines with TG 51 non compliant beams. Acrylic OSLD miniphantoms are used for the CyberKnife. Special TLD phantoms are used for TomoTherapy and GammaKnife machines to accommodate the specific geometry of each machine. These remote audit tools are sent to institutions to be irradiated and returned to IROC-H for analysis. Results: The average IROC-H/institution ratios for 480 GammaKnife, 660 CyberKnife and 907 rotational TomoTherapy beams are 1.000±0.021, 1.008±0.019, 0.974±0.023, respectively. In the particular case of TomoTherapy, the overall ratio is 0.977±0.022 for HD units. The standard deviations of all results are consistent with values determined for TG 51 compliant photon beams. These ratios have shown some changes compared to values presented in 2008. The GammaKnife results were corrected by an experimentally determined scatter factor of 1.025 in 2013. The TomoTherapy helical beam results are now from a rotational beam whereas in 2008 the results were from a static beam. The decision to change modality was based on recommendations from the users. Conclusion: External audits of beam outputs is a valuable tool to confirm the calibrations of photon beams regardless of whether the machine is TG 51 or TG 51 non compliant. The difference found for TomoTherapy units is under investigation. This investigation was supported by IROC grant CA180803 awarded by the NCI.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2014

Neutronics for diagnostics in ITER - Paper 121

ITER, currently under construction in the south of France, will be the first tokamak producing large amounts of neutrons and having a fusion energy gain factor, which is the ratio between the fusion power produced and the input power necessary to maintain the plasma, greater than 10. ITER will be able to generate 500 MW of fusion power, from which 400 MW are 14.1 MeV neutrons from the Deuterium-Tritium fusion reactions in the thermonuclear plasma. For such nuclear machine neutronics are of paramount importance, being a design driver for many of its components, in particular its Ports and its numerous Diagnostics, for which the influence of neutronics is especially important in three main areas: - Shielding: there is a strong need to provide nuclear shielding to areas where human access is foreseen for maintenance. Also, to minimize any impact in the primary shielding components (blankets) in order to avoid significant increases in the heating of the Vacuum Vessel (VV) and Superconductive Coils. - Sensors and diagnostic components: most of the diagnostics have sensitive components that need to be shielded against radiation to survive. Efficient shielding is not always possible, and the radiation levels are usually the decisive factor for selecting the most appropriate component. - Neutron measurements: a large set of neutron diagnostics will be implemented to measure the neutron production and its emissivity profile. Although the scale and performance of ITER will not be very far from a commercial reactor, it is ultimately an experiment to improve the understanding of reactor-grade tokamak plasmas and their effect on plasma facing components. To accomplish this task, large efforts are being dedicated to design over 45 different diagnostics systems. This paper gives an overview of the main activities related to neutronics within the ITER Diagnostics Division. They include the neutron streaming problems in the Ports and the common neutronics issues of the diagnostics housed inside them, as well as the challenges of the in vessel diagnostics. Also the special calculations needed for the neutron diagnostics. (authors)

61 RADIATION PROTECTION AND DOSIMETRY↗

DOE OSTI2021

Identifying schools at high-risk for elevated lead in drinking water using only publicly available data

Estimating the risk of lead contamination of schools' drinking water at the State level is a complex, important, and unexplored challenge. Variable water quality among water systems and changes in water chemistry during distribution affect lead dissolution rates from pipes and fittings. In addition, the locations of lead-bearing plumbing materials are uncertain. We tested the capability of six machine learning models to predict the likelihood of lead contamination of drinking water at the schools' taps using only publicly available datasets. The predictive features used in the models correspond to those with a proven correlation to the dominant, but commonly unavailable, factors that govern lead leaching: the presence of lead-bearing plumbing materials and water quality conducive to lead corrosion. By combining water chemistry data from public reports, socioeconomic information from the US census, and spatial features using Geographic Information Systems, we trained and tested models to estimate the likelihood of lead contaminated tap water in over 8,000 schools across California and Massachusetts. Our best-performing model was a Random Forest, with a 10-fold cross validation score of 0.88 for Massachusetts and 0.78 for California using the average Area Under the Receiver Operating Characteristic Curve (ROC AUC) metric. The model was then used to assign a lead leaching risk category to half of the schools across California (the other half was used for training). There was good agreement between the modeled risk categories and the actual lead leaching outcomes for every school; however, the model overestimated the lead leaching risk in up to 17% of the schools. This model is the first of its kind to offer a tool to predict the risk of lead leaching in schools at the State level. Further use of this model can help deploy limited resources more effectively to prevent childhood lead exposure from school drinking water.

63 RADIATION, THERMAL, AND OTHER ENVIRON. POLLUTAN↗

DOE OSTI2022

An autonomous robot for shell and tube heat exchanger inspection

Shell and tube heat exchangers (STHEs) are critical to energy conversion efficiency of power plants. Eddy current examination is a way to evaluate working conditions of these tubes. However, the current testing apparatus requires human to manually insert an eddy current testing (ECT) probe into and extract it out of individual tubes, and meanwhile monitor measurement results for diagnosis. It is a time-consuming and labor-intensive procedure even for an experienced technician. To tackle this challenge, in this study, we developed a robot enabled ECT system for autonomous inspection of STHEs. The robotic platform employs Mecanum wheeled chassis for high mobility, machine vision to locate tube bundle and tube inlets, a rotational Cartesian mechanism to operate at planes with all possible inclinations, and a task-specific mechanism for ECT probe delivery. Machine vision locates tube bundle and tube inlets by an April tag detection algorithm and a Circle Hough Transform algorithm, respectively. Assisted by a guiding cone, the ECT probe is continuously fed into the tubes with a fill factor of 0.819. In conclusion, during this process, the eddy current data are automatically collected and real-time analyzed by convolutional neural networks, showing accuracy of nearly 100% for identifying defective and nondefective tubes and 85% for four types of defective tubes and nondefective tubes.

42 ENGINEERING↗

DOE OSTI2025

Evolution and Degradation Patterns of Electrochemical Cells Based on the Analysis of Interfacial Phenomena at Li Metal Anode/Electrolyte Interfaces

In this work, we report the results of a theoretical–computational analysis of the solid electrolyte interphase (SEI) growth and degradation dynamics occurring in lithium metal batteries during cycling. We use ab initio-kinetic Monte Carlo simulations to generate a synthetic data set, which is analyzed by machine learning methods. We aim to determine: (i) how modifications in interfacial interaction energies between solid electrolyte interphase (SEI) blocks and between Li ions and SEI facets impact the Coulombic efficiency (CE) of the battery and (ii) what factors, including reactions, microscopic transport, and other interfacial events, may lead to cell performance “failure” during prolonged charge and discharge cycles, signaled as a sharp decay in the CE over cycling. The demonstration of our approach is done on a cell including a Li metal surface interfacing with a previously introduced state-of-the-art electrolyte, and the idea can be applied to any electrochemical system. Outcomes include the identification of the leading chemical, physical, and structural variables causing cell failure and relating them to the electrolyte formulation, thus paving the way to future more refined analysis and electrolyte design.

batteries↗

DOE OSTI2024

Dynamic Multiplexed Control and Modeling of Optogenetic Systems Using the High-Throughput Optogenetic Platform, Lustro

The ability to control cellular processes using optogenetics is inducer-limited, with most optogenetic systems responding to blue light. To address this limitation, we leverage an integrated framework combining Lustro, a powerful high-throughput optogenetics platform, and machine learning tools to enable multiplexed control over blue light-sensitive optogenetic systems. Specifically, we identify light induction conditions for sequential activation as well as preferential activation and switching between pairs of light-sensitive split transcription factors in the budding yeast, Saccharomyces cerevisiae. We use the high-throughput data generated from Lustro to build a Bayesian optimization framework that incorporates data-driven learning, uncertainty quantification, and experimental design to enable the prediction of system behavior and the identification of optimal conditions for multiplexed control. This work lays the foundation for designing more advanced synthetic biological circuits incorporating optogenetics, where multiple circuit components can be controlled using designer light induction programs, with broad implications for biotechnology and bioengineering.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2019

Direct simulation Monte Carlo on petaflop supercomputers and beyond

The gold-standard definition of the Direct Simulation Monte Carlo (DSMC) method is given in the 1994 book by Bird [Molecular Gas Dynamics and the Direct Simulation of Gas Flows (Clarendon Press, Oxford, UK, 1994)], which refined his pioneering earlier papers in which he first formulated the method. In the intervening 25 years, DSMC has become the method of choice for modeling rarefied gas dynamics in a variety of scenarios. The chief concern to applying DSMC to more dense or even continuum flows is its computational expense compared to continuum computational fluid dynamics methods. The dramatic (nearly billion-fold) increase in speed of the largest supercomputers over the last 30 years has thus been a key enabling factor in using DSMC to model a richer variety of flows, due to the method’s inherent parallelism. We have developed the open-source SPARTA DSMC code with the goal of running DSMC efficiently on the largest machines, both current and future. It is largely an implementation of Bird’s 1994 formulation. In this work, we describe algorithms used in SPARTA to enable DSMC to operate in parallel at the scale of many billions of particles or grid cells, or with billions of surface elements. We highlight a few examples of the kinds of fundamental physics questions and engineering applications that DSMC can address at these scales.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Deterministic High-Fidelity Neutronics Simulation of Pebble Bed Reactors Using Pebble Tracking Transport

The pebble tracking transport (PTT) algorithm offers a high-fidelity deterministic approach for neutron transport for pebble bed reactors (PBRs). This approach requires the mesh for the active-core region to consist exclusively of tetrahedral elements, where each node in the pebble-packing region represents a pebble centroid. This paper investigates the application of PTT for full-scale PBRs, considering both the isothermal and the temperature-dependent core conditions. Macroscopic cross sections are generated using Serpent 2 full-core eigenvalue simulations where pebbles are grouped into disjoint subsets using machine learning. To minimize the need for individual cross-section sets for each pebble in the core, K-means clustering is used to group pebbles by temperature and neutronic environment parameters. Here, we compare the multiplication factor and power rate distributions between PTT simulations using the Griffin reactor physics software and reference solutions from Serpent 2. Our analysis shows that a full-core, high-fidelity PTT calculation produces accurate results with minimal local (pebblewise) errors. Additionally, timing results indicate that PTT simulations converge rapidly on modern supercomputing platforms.

Griffin↗

DOE OSTI2026

Recent advances in plasma control and physics research in the Large Helical Device

The Large Helical Device (LHD), the largest superconducting helical system in the world, is equipped with advanced heating and diagnostic tools, facilitating plasma control and physics research. Data assimilation was employed for electron temperature control using a real-time Thomson scattering system and real time prediction code. A virtual LHD environment enabled visualization of escaping high-energy tritium ions and demonstrated that these ions impact the rear side of the divertor plate. Pioneering results crucial to plasma control have also been achieved. Real-time wall conditioning using Lithium granule dropping improved bulk ion energy and particle transport while simultaneously enhancing the heavy impurity transport. Progress has also been made in the investigation of turbulence-driven transport. At the confinement bifurcation, ion-scale turbulence decreased, while electron-scale turbulence increased. A change in the anisotropy of turbulent eddies was also observed at the confinement bifurcation. Coexistence of local and non-local turbulence was identified in electron-scale turbulence. Non-local turbulence exhibited the rapid spatial propagation of perturbations throughout the plasma, while local turbulence followed the temperature gradient. A transition between drift-wave turbulence and magnetohydrodynamics (MHD) turbulence was observed with the turbulence minimized at the transition condition. Machine learning analysis was employed to evaluate the temperate and density conditions of this turbulence transition. Then, real-time control of fueling and heating was applied to maintain the turbulence transition condition, improving the energy confinement enhancement factor by 20%. In addition, evidence was obtained for collisionless ion heating by energetic-ion-driven geodesic acoustic modes and MHD bursts. These achievements represent unique contributions to the development of fusion reactors.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2021

Discovery of two bright high-redshift gravitationally lensed quasars revealed by Gaia

We present the discovery and preliminary characterisation of two high-redshift gravitationally lensed quasar systems in Gaia Data Release 2 (DR2). Candidates with multiple close-separation Gaia detections and quasar-like colours in WISE, Pan-STARRS, and DES are selected for follow-up spectroscopy with the New Technology Telescope. We confirm DES J215028.71-465251.3 as a $z$ = 4.130 ± 0.006 asymmetric, doubly imaged lensed quasar system and model the lensing mass distribution as a singular isothermal sphere. The system has an Einstein radius of 1.202 ± 0.005 arcsec and a predicted time delay of ~122.0 d between the quasar images, assuming a lensing galaxy redshift of $z$ = 0.5, making this a priority system for future optical monitoring. We confirm PS J042913.17+142840.9 as a $z$ = 3.866 ± 0.003 four-image quasar system in a cusp configuration, lensed by two foreground galaxies. The system is well modelled using a singular isothermal ellipsoid for the primary lens and a singular isothermal sphere for the secondary lens with Einstein radii 0.704 ± 0.006 and 0.241 ± 0.030 arcsec, respectively. A maximum predicted time delay of 9.6 d is calculated, assuming lensing galaxy redshifts of $z$ = 1.0. Furthermore, PS J042913.17+142840.9 exhibits a large flux ratio anomaly, up to a factor of 2.66 ± 0.37 in i band, that varies across optical and near-infrared wavelengths. We discuss LSST and its implications for future high-redshift lens searches and outline an extension to the search using supervised machine learning techniques.

Astronomy & Astrophysics↗

DOE OSTI2026

Identifying Anomalous DESI Galaxy Spectra with a Variational Autoencoder

The tens of millions of spectra being captured by the Dark Energy Spectroscopic Instrument (DESI) provide tremendous discovery potential. In this work we show how Machine Learning, in particular Variational Autoencoders (VAE), can detect anomalies in a sample of approximately 200,000 DESI spectra comprising galaxies, quasars and stars. We demonstrate that the VAE can compress the dimensionality of a spectrum by a factor of 100, while still retaining enough information to accurately reconstruct spectral features. We then detect anomalous spectra as those with high reconstruction error and those which are isolated in the VAE latent representation. The anomalies identified fall into two categories: spectra with artefacts and spectra with unique physical features. Awareness of the former can help to improve the DESI spectroscopic pipeline; whilst the latter can lead to the identification of new and unusual objects. To further curate the list of outliers, we use the Astronomaly package which employs Active Learning to provide personalised outlier recommendations for visual inspection. In this work we also explore the VAE latent space, finding that different object classes and subclasses are separated despite being unlabelled. We demonstrate the interpretability of this latent space by identifying tracks within it that correspond to various spectral characteristics. For example, we find tracks that correspond to increasing star formation and increase in broad emission lines along the Balmer series. In upcoming work we hope to apply the methods presented here to search for both systematics and astrophysically interesting objects in much larger datasets of DESI spectra.

Nicolaou, C. [University Coll. London] (ORCID:0000↗

DOE OSTI2022

EIC Beam Dynamics Challenges

The Electron Ion Collider aims to produce luminosities of 10³⁴ cm⁻²s⁻¹ . The machine will operate over a broad range of collision energies with highly polarized beams. The coexistence of highly radiative electrons and nonradiative ions produce a host of unique effects. Strong hadron cooling will be employed for the final factor of 3 luminosity boost.

43 PARTICLE ACCELERATORS↗

DOE OSTI2025

A Near-Real-Time Model for Predicting Electricity Disruptions in Texas During Winter Storms

There has been an increase in extreme weather events, posing a threat to power grid systems, potentially influenced by factors such as population growth, changes in ecosystems, land cover, and land use in the service area, as well as the growth of certain vegetation types. This research seeks to develop a predictive model to mitigate potential damages caused by future winter storms. This research utilizes the Light Gradient Boosting Machine (LightGBM), incorporating the number of power outages experienced at the county level, geographic details, weather information, and lagged outage and lagged weather data. The developed models were broadly divided into two groups, with six models in each group - one group without optimization and another with optimization, totaling 12 trained models. For model optimization, Bayesian optimization was employed using Root Mean Squared Error (RMSE) as the objective function. In results, when comparing Group 2 (the optimized group) with Group 1 (the non-optimized group), it was found that optimization did not always lead to a reduction in RMSE and Mean Absolute Error (MAE). However, in terms of Mean Directional Accuracy (MDA), while all results in Group 1 were below the baseline accuracy of 0.33, all results in Group 2 exceeded 0.33, with some cases showing an increase of more than three times the baseline. The results indicated that, in the optimized model group, Population and Pressure were the most influential factors when using current weather data and geographical information. When using lagged data, lagged recorded outages and lagged Pressure emerged as the most significant factors. Among the 12 developed models, the L-1-2-O model showed the lowest RMSE and MAE, as well as the highest accuracy, with values of 390.62 households and 168.13 households, respectively. To normalize the RMSE and MAE values, each metric was divided by the average number of households among the counties in Texas. For the L-1-2-O model, the scaled RMSE was 0.88% and the scaled MAE was 0.38%. In terms of MDA, which indicates the accuracy of the prediction direction, the L-1-O model achieved the highest score of 0.41. Although this study focused on Texas, which suffered the greatest impact from the winter storms in 2021, with additional validation, the methodology used in this research could be applied to other regions.

Lee, Jangjae [Texas A & M Univ., College Station, ↗

DOE OSTI2024

Side-by-Side Comparison of Subhourly Clipping Models

Over the past several years there have been numerous attempts at quantifying the inherent power clipping of inverters due to subhourly irradiance variability that is not captured in hourly PV performance models. Different models have been proposed to correct for these clipping losses in PV performance estimates, including matrix lookup models, distribution modeling of the PV power performance within a given hour, and machine learning methods. To date, there have been few comprehensive quantitative comparisons of these inverter clipping correction modeling approaches to evaluate the effectiveness of these approaches in predicting the actual behavior of PV system inverter clipping. In this study, we perform such a comparison, evaluating the Allen and Walker correction loss modeling approaches recently implemented in the System Advisor Model (SAM) against clipping losses modeled with 1-minute climate data. These comparisons were performed across a variety of climate locations and inverter loading ratios to thoroughly analyze the effectiveness of these modeling approaches relative to each other. Results from this analysis reveal that both clipping correction approaches improve annual energy accuracy to within 2% of 1-minute modeled energy yield. The two models predict annual clipping loss more accurately than simple hourly power limit clipping, with the Allen method typically being slightly more accurate at typical ILR values and the Walker method often being slightly more accurate at high ILR values The models can improve accuracy over the status quo clipping approach up to 3 percentage points in systems with ILR of 2.0, showing the importance of this modeling factor in energy yield estimates.

accuracy↗

DOE OSTI2024

Teaching Freight Mode Choice Models New Tricks Using Interpretable Machine Learning Methods

Understanding and forecasting the intricate freight mode choice behavior under various industry, policy, and technology contexts is essential in freight planning and policymaking. Numerous models have been developed in prior studies to provide insights into freight mode selection, the majority of which use discrete choice models such as multinomial logit (MNL) models. However, logit models often rely on linear specifications of independent variables, despite potential nonlinear relationships in the data. Moreover, there often lacks a heuristic and efficient approach to identify such complex relationships to define the logit model specifications. To fill this gap, we developed an MNL model for freight mode choice using the insights from state-of-the- art machine learning (ML) models. ML models can capture the nonlinear nature of the complex decision-making process, and recent advances in 'explainable AI' have greatly improved their interpretability. The interpretable ML methods help enhance the performance of MNL models and advance knowledge of freight mode choice. Specifically, the influential factors and their relationship with individual modes are identified using SHapley Additive exPlanations (SHAP) to improve the MNL's performance. The workflow is demonstrated in a case study of Austin, Texas, and the SHAP results reveal multiple nonlinear relationships predicted by ML models. Incorporating those relationships into MNL model specifications improves the interpretability and accuracy of the MNL model compared to a conventional MNL model. Findings from this study can be used to guide freight planning and inform policymakers and practitioners on how key factors affect freight decision-making.

ADVANCED PROPULSION SYSTEMS,MATHEMATICS AND COMPUT↗

DOE OSTI2018

Modeling the Impact of Baryons on Subhalo Populations with Machine Learning

Here, we identify subhalos in dark matter–only (DMO) zoom-in simulations that are likely to be disrupted due to baryonic effects by using a random forest classifier trained on two hydrodynamic simulations of Milky Way (MW)–mass host halos from the Latte suite of the Feedback in Realistic Environments (FIRE) project. We train our classifier using five properties of each disrupted and surviving subhalo: pericentric distance and scale factor at first pericentric passage after accretion and scale factor, virial mass, and maximum circular velocity at accretion. Our five-property classifier identifies disrupted subhalos in the FIRE simulations with an 85% out-of-bag classification score. We predict surviving subhalo populations in DMO simulations of the FIRE host halos, finding excellent agreement with the hydrodynamic results; in particular, our classifier outperforms DMO zoom-in simulations that include the gravitational potential of the central galactic disk in each hydrodynamic simulation, indicating that it captures both the dynamical effects of a central disk and additional baryonic physics. We also predict surviving subhalo populations for a suite of DMO zoom-in simulations of MW-mass host halos, finding that baryons impact each system consistently and that the predicted amount of subhalo disruption is larger than the host-to-host scatter among the subhalo populations. Although the small size and specific baryonic physics prescription of our training set limits the generality of our results, our work suggests that machine-learning classification algorithms trained on hydrodynamic zoom-in simulations can efficiently predict realistic subhalo populations.

79 ASTRONOMY AND ASTROPHYSICS↗