Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “deep potential”

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.

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

At least 1,675 records · Page 93
All sourcesNASA NTRSDOE OSTI
DOE OSTI2013

Element partitioning in combustion- and gasification-based waste-to-energy units

Highlights: ► Element partitioning of waste-to-energy units by means of a substance flow analysis. ► A comparison between moving grate combustors and high temperature gasifiers. ► Classification of key elements according to their behavior during WtE processes. ► Slags and metals from waste gasifiers are completely and immediately recyclable. ► Potential reduction of amounts of solid residue to be sent to landfill disposal. - Abstract: A critical comparison between combustion- and gasification-based waste-to-energy systems needs a deep knowledge of the mass flows of materials and elements inside and throughout the units. The study collected and processed data from several moving grate conventional incinerators and high-temperature shaft gasifiers with direct melting, which are in operation worldwide. A material and substance flow analysis was then developed to systematically assess the flows and stocks of materials and elements within each waste-to-energy unit, by connecting the sources, pathways, and intermediate and final sinks of each species. The patterns of key elements, such as carbon, chloride and heavy metals, in the different solid and gaseous output streams of the two compared processes have been then defined. The combination of partitioning coefficients with the mass balances on atomic species and results of mineralogical characterization from recent literatures was used to estimate a composition of bottom ashes and slags from the two types of waste-to-energy technologies. The results also allow to quantify some of the performance parameters of the units and, in particular, the potential reduction of the amount of solid residues to be sent to final disposal.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
NASA NTRS2004

Characterization of Transport Errors in Chemical Forecasts from a Global Tropospheric Chemical Transport Model

We propose a new methodology to characterize errors in the representation of transport processes in chemical transport models. We constrain the evaluation of a global three-dimensional chemical transport model (GEOS-CHEM) with an extended dataset of carbon monoxide (CO) concentrations obtained during the Transport and Chemical Evolution over the Pacific (TRACE-P) aircraft campaign. The TRACEP mission took place over the western Pacific, a region frequently impacted by continental outflow associated with different synoptic-scale weather systems (such as cold fronts) and deep convection, and thus provides a valuable dataset. for our analysis. Model simulations using both forecast and assimilated meteorology are examined. Background CO concentrations are computed as a function of latitude and altitude and subsequently subtracted from both the observed and the model datasets to focus on the ability of the model to simulate variability on a synoptic scale. Different sampling strategies (i.e., spatial displacement and smoothing) are applied along the flight tracks to search for systematic model biases. Statistical quantities such as correlation coefficient and centered root-mean-square difference are computed between the simulated and the observed fields and are further inter-compared using Taylor diagrams. We find no systematic bias in the model for the TRACE-P region when we consider the entire dataset (i.e., from the surface to 12 km ). This result indicates that the transport error in our model is globally unbiased, which has important implications for using the model to conduct inverse modeling studies. Using the First-Look assimilated meteorology only provides little improvement of the correlation, in comparison with the forecast meteorology. These general statements can be refined when the entire dataset is divided into different vertical domains, i.e., the lower troposphere (less than 2 km), the middle troposphere (2-6 km), and the upper troposphere (greater than 6 km). The best agreement between the observations and the model is found in the lower and middle troposphere. Downward displacements in the lower troposphere provide a better fit with the observed value, which could indicate a problem in the representation of boundary layer height in the model. Significant improvement is also found for downward and southward displacements in the upper troposphere. There are several potential sources of errors in our simulation of the continental outflow in the upper troposphere which could lead to such biases, including the location and/or the strength of deep convective cells as well as that of wildfires in Southeast Asia.

Bey, I.↗
DOE OSTI2020

High precision control and deep learning-based corn stand counting algorithms for agricultural robot

This paper presents high precision control and deep learning-based corn stand counting algorithms for a low-cost, ultra-compact 3D printed and autonomous field robot for agricultural operations. Currently, plant traits, such as emergence rate, biomass, vigor, and stand counting, are measured manually. This is highly labor-intensive and prone to errors. The robot, termed TerraSentia, is designed to automate the measurement of plant traits for efficient phenotyping as an alternative to manual measurements. In this paper, we formulate a Nonlinear Moving Horizon Estimator that identifies key terrain parameters using onboard robot sensors and a learning-based Nonlinear Model Predictive Control that ensures high precision path tracking in the presence of unknown wheel-terrain interaction. Moreover, we develop a machine vision algorithm designed to enable an ultra-compact ground robot to count corn stands by driving through the fields autonomously. The algorithm leverages a deep network to detect corn plants in images, and a visual tracking model to re-identify detected objects at different time steps. We collected data from 53 corn plots in various fields for corn plants around 14 days after emergence (stage V3 - V4). The robot predictions have agreed well with the ground truth with C robot =1.02×C human -0.86 and a correlation coefficient R=0.96. The mean relative error given by the algorithm is -3.78%, and the standard deviation is 6.76%. These results indicate a first and significant step towards autonomous robot-based real-time phenotyping using low-cost, ultra-compact ground robots for corn and potentially other crops.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2013

Bay Ridge Gardens - Mixed-Humid Affordable Multifamily Housing Deep Energy Retrofit

Under this project, Newport Partners (as part of the BA-PIRC research team) evaluated the installation, measured performance, and cost-effectiveness of efficiency upgrade measures for a tenant-in-place DER at the Bay Ridge multifamily (MF) development in Annapolis, Maryland. The design and construction phase of the Bay Ridge project was completed in August 2012. This report summarizes system commissioning, short-term test results, utility bill data analysis, and analysis of real-time data collected over a one-year period after the retrofit was complete. The Bay Ridge project is comprised of a 'base scope' retrofit which was estimated to achieve a 30%+ savings (relative to pre-retrofit) on 186 apartments, and a 'DER scope' which was estimated to achieve 50% savings (relative to pre-retrofit) on a 12-unit building. The base scope was applied to the entire apartment complex, except for one 12-unit building which underwent the DER scope. A wide range of efficiency measures was applied to pursue this savings target for the DER building, including improvements/replacements of mechanical equipment and distribution systems, appliances, lighting and lighting controls, the building envelope, hot water conservation measures, and resident education. The results of this research build upon the current body of knowledge of multifamily retrofits. Towards this end, the research team has collected and generated data on the selection of measures, their estimated performance, their measured performance, and risk factors and their impact on potential measures.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2021

Manipulating meso-scale solvent structure from Pd nanoparticle deposits in deep eutectic solvents

Deep Eutectic Solvents (DESs) are complex solutions that present unique challenges compared to traditional electrolytes. Unlike most electrolytes and ionic liquids, DESs have delicate hydrogen bond networks that are responsible for their highly sensitive compositional dependence on the melting point. Prior work has demonstrated unique nanoscale structure both experimentally and theoretically that brings both challenges and opportunities to their adoption in traditional electrochemical processes. In this study, we use in-situ sample-rotated ultra-small angle X-ray scattering to resolve the near-interface solvent structure after electrodepositing Pd nanoparticles onto a glassy carbon electrode in choline chloride:urea and choline chloride:ethylene glycol DESs. Our results indicate that solvent structure can be observed on the meso-scale in the choline chloride:urea and choline chloride:ethylene glycol systems. Importantly, this extended solvent structure peaks at -0.5 V (vs. Ag/AgCl) and decreases slightly with higher overpotentials during Pd electrodeposition. Experimentally, the nature of this structure is more pronounced in the ethylene glycol system, as evidenced by both the X-ray scattering as well as the electrochemical impedance spectroscopy (EIS). Here, molecular dynamics simulations and dipolar orientation analysis reveal that chloride delocalization near the Pd interface and long-range interactions between the choline and each hydrogen bond donor are very different and qualitatively consistent with the experimental data. These results show how the long-range solvent-deposit interactions can be tuned by changing the hydrogen bond donor in the DES and the applied potential.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2018

Erroneous Attribution of Deep Convective Invigoration to Aerosol Concentration

Abstract Contiguous time–height cloud objects at the Department of Energy Atmospheric Radiation Measurement Southern Great Plains (SGP) site are matched with surface condensation nuclei (CN) concentrations and retrieved thermodynamic and kinematic vertical profiles for warm-cloud-base, cold-cloud-top systems in convectively unstable environments. Statistical analyses show that previously published conclusions that increasing CN concentrations cause a decrease in minimum cloud-top temperature (CTT) at the SGP site through the aerosol convective invigoration effect are unfounded. The CN–CTT relationship is statistically insignificant, while correlations between convective available potential energy (CAPE), level of neutral buoyancy (LNB), and CN concentration account for most of the change in the CN–CTT positive correlation. Removal of clouds with minimum CTTs > −36°C from the analysis eliminates the CN–CTT correlation. Composited dirty conditions at the SGP have ~1°C-warmer low levels and ~1°C-cooler upper levels than clean conditions. This correlation between aerosol concentrations and thermodynamic profiles may be caused by an increase in regional rainfall preceding deep convective conditions as CN concentration decreases. Increased rainfall can be expected to increase wet deposition of aerosols, cool low-level temperatures, and warm upper-level temperatures. The masking of a potential aerosol effect by such small thermodynamic changes implies that the strategy of analyzing subsets of aerosol data by binned meteorological factor values is not a valid method for discerning an aerosol effect in some situations. These findings highlight the need for more careful, detailed, and strategic observations to confidently isolate and quantify an aerosol deep convective invigoration effect.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

In situ compression artifact removal in scientific data using deep transfer learning and experience replay

The massive amount of data produced during simulation on high-performance computers has grown exponentially over the past decade, exacerbating the need for streaming compression and decompression methods for efficient storage and transfer of this data---key to realizing the full potential of large-scale computational science. Lossy compression approaches such as JPEG when applied to scientific simulation data realized as a stream of images can achieve good compression rates but at the cost of introducing compression artifacts and loss of information. This paper develops a unified framework for in situ compression artifact removal in which the fully convolutional neural network architectures are combined with scalable training, transfer learning, and experience replay to achieve superior accuracy and efficiency while significantly decreasing the storage footprint as compared with the traditional optimization-based approaches. We demonstrate the proposed approach and compare it with compressed sensing postprocessing and other baseline deep learning models using climate simulations and nuclear reactor simulations, both of which are driven by hyperbolic partial differential equations. Our approach when applied to remove the compression artifacts on the JPEG-compressed nuclear reactor simulation data (using a transfer-trained model that was pretrained on the climate simulation data and updated incrementally as the nuclear reactor simulation progressed), achieved a significant improvement---mean peak signal-to-noise ratio of 42.438 as compared with 27.725 obtained with the compressed sensing approach.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Engineering shallow and deep level defects in $\kappa$-Ga 2 O 3 thin films: comparing metal-organic vapour phase epitaxy to molecular beam epitaxy and the effect of annealing treatments

Orthorhombic gallium oxide (κ-Ga 2 O 3 ) is an ultra-wide bandgap semiconductor with great potential in new generation electronics. Its application is hindered at present by the limited physical understanding of the relationship between synthesis and functional properties. This work discusses the effects of growth method (metal-organic vapour phase epitaxy and molecular beam epitaxy) as well as annealing treatments in different atmospheres (O 2 , H 2 ) on point defects in κ-Ga 2 O 3 layers epitaxially grown on c-plane sapphire. Comprehensive experimental characterization by X-ray diffraction, photo current-as well as photoluminescence excitation spectroscopy, and X-ray photo electron spectroscopy is combined with first principles calculations of the point defects’ formation and complex-dissociation energies. We demonstrate that for κ-Ga 2 O 3 the concentration of shallow and deep level defects can be sensitively controlled through annealing treatments at temperatures (T = 500 °C) well below the thermal stability threshold of this polymorph. In particular, our results suggest that hydrogen-related defects (e.g., H-interstitials, Ga-vacancies—H complexes) play a key role in this process. While we provide direct exemplary implications of our results for the performances of κ-Ga 2 O 3 based photodetectors, these findings are predicted to impact further application fields of κ-Ga 2 O 3 , such as high electron mobility transistors or memory devices.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Long-length Scintillating Fibers for Nuclear Geological Repositories

This research focuses on the usage of plastic scintillating fibers as ruggedized detectors in long-length sections coupled to long-length standard optical fiber segments for the purpose of radiation detection and monitoring. This technology addresses a need for monitoring radiation in difficult-to-access or inaccessible areas, such as monitoring: radioactive waste within geological repositories, nuclear material storage facilities, spent fuel dry storage areas, inaccessible areas within nuclear facilities including reactors and reprocessing facilities, decommissioning of nuclear facilities, and other similar areas. The fiber can be placed along the tunnels of a geological repository to monitor the radiation field along its length. This can detect an increase of radiation due to movement, leakage, or unexpected opening of the containers, as well as decrease in the radiation due to removal of material. These capabilities have been demonstrated at INL’s Radioactive Scrap Waste Facility by successfully detecting the removal of an interim storage containers from the storage area. The poster focuses on explaining the technology and its capability in monitoring deep geological repositories (DGRs), especially in the domain of safety and international safeguards. As more countries consider the development of DGRs, technologies such as the one being presented can help establish continuity of knowledge of the material, maintain public confidence, and ensure safe operation of the facility. Furthermore, the poster looks at the future steps of the technology and its potential role in the international safeguards regime.

98 - NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL↗
DOE OSTI2019

Absence of hindrance in a microscopic 12 C + 12 C fusion study

Studies of low-energy fusion of light nuclei are important in astrophysical modeling, with small variations in reaction rates having a large impact on nucleosynthesis yields. Due to the lack of experimental data at astrophysical energies, extrapolation and microscopic methods are needed to model fusion probabilities. To investigate deep sub-barrier 12 C+ 12 C fusion cross sections and establish trends for the S factor. Microscopic methods based on static Hartree-Fock and time-dependent Hartree-Fock (TDHF) mean-field theory are used to obtain 12 C+ 12 C ion-ion fusion potentials. Fusion cross sections and astrophysical S factors are then calculated using the incoming wave boundary condition method. Both density-constrained frozen Hartree-Fock (DCFHF) and density-constrained TDHF (DC-TDHF) predict a rising S factor at low energies, with DC-TDHF predicting a slight damping in the deep sub-barrier region (≈1 MeV). Comparison between DC-TDHF calculations and maximum experimental cross sections in the resonance peaks are good. However, the discrepancy in experimental low-energy results inhibits interpretation of the trend. Using the fully microscopic DCFHF and DC-TDHF methods, no S factor maximum is observed in the 12 C+ 12 C fusion reaction. In addition, no extreme sub-barrier hindrance is predicted at low energies. Furthermore, the development of a microscopic theory of fusion including resonance effects, as well as further experiments at lower energies must be done before the deep sub-barrier behavior of the reaction can be established.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Updraft Width Modulates Ambient Atmospheric Controls on Convective Cloud Depth

Abstract The depth of convective clouds affects vertical transport of atmospheric constituents, influencing downstream weather and climate. Atmospheric controls on the maximum depth reached by moist convection are investigated with radar‐tracked convective cells tagged with sounding‐derived atmospheric parameters from a field campaign in central Argentina. Regression analyses show that narrow (<12‐km diameter) and wide (>16‐km diameter) cell depths respond to disparate factors, where cell areas are defined using composite reflectivity signatures. Undiluted lifted parcel indices including convective available potential energy (CAPE) and level of neutral buoyancy (LNB) are top predictors of wide cell maximum depth while mid‐tropospheric relative humidity is the top predictor of narrow cell maximum depth. Because narrow cells are more numerous than wide cells, the overall outcome of the full cell population does not strongly correlate with CAPE and LNB conditions. Tracked cells and atmospheric conditions in a simulation with 3‐km grid spacing covering the field campaign produce similar results to those observed. Narrow cells that are relatively deep have a cooler and moister mid‐troposphere with weaker free tropospheric subsidence, while relatively deep wide cells have much warmer and moister lower tropospheric conditions. These atmospheric differences are present 1 hr before cell initiation at both a fixed observing site and variable cell initiation locations. Simulated narrow cell maximum equivalent potential temperature decreases with height at a rate similar to the ambient vertical gradient, causing these cells to fall short of their LNB and supporting the view that entrainment‐driven dilution is a dominant control on their depth.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Protein sequence design with a learned potential

The task of protein sequence design is central to nearly all rational protein engineering problems, and enormous effort has gone into the development of energy functions to guide design. Here, we investigate the capability of a deep neural network model to automate design of sequences onto protein backbones, having learned directly from crystal structure data and without any human-specified priors. The model generalizes to native topologies not seen during training, producing experimentally stable designs. We evaluate the generalizability of our method to a de novo TIM-barrel scaffold. The model produces novel sequences, and high-resolution crystal structures of two designs show excellent agreement with in silico models. Our findings demonstrate the tractability of an entirely learned method for protein sequence design.

59 BASIC BIOLOGICAL SCIENCES↗
NASA NTRSDate not supplied

What Might the End Permian Tell Us About Earth’s Deep Future Climate

Earth has supported life for over 3 billion years and is expected to do so for ~3 billion years more [1]. Of particular interest to humans is the relatively shorter time during which multicellular life evolves and thrives. This may be more critical to understanding the potential for complex ecosystems like our own developing in other star systems. We have explored the conditions on Earth at either end of a 500 Myr span from the present day. These time periods are both associated with supercontinents and may be considered more habitable than our modern world. The End Permian (~250 Ma) mass extinction event (eliminating ~90% of life on land and ~80% in the ocean) played a critical role in the evolution of life on Earth when land was amalgamated into a supercontinent called Pangea, and generally high greenhouse gas (GHG) concentrations provided a counter to an insolation ~2% less than today. In the deep future (+250 Myr), a hypothetical low latitude supercontinent state called Aurica [2] is projected to still support life as we know it despite an insolation 2% higher than today [1;Fig 5].

End permian↗
DOE OSTI2023

Transferable Adversarial Attack on 3D Object Tracking in Point Cloud

3D point cloud object tracking has recently witnessed considerable progress relying on deep learning. Such progress, however, mainly focuses on improving tracking accuracy. The risk, especially considering that deep neural network is vulnerable to adversarial perturbations, of a tracker being attacked is often neglected and rarely explored. In order to attract attentions to this potential risk and facilitate the study of robustness in point cloud tracking, we introduce a novel transferable attack network (TAN) to deceive 3D point cloud tracking. Specifically, TAN consists of a 3D adversarial generator, which is trained with a carefully designed multi-fold drift (MFD) loss. The MFD loss considers three common grounds, including classification, intermediate feature and angle drifts, across different 3D point cloud tracking frameworks for perturbation generation, leading to high transferability of TAN for attack. In our extensive experiments, we demonstrate the proposed TAN is able to not only drastically degrade the victim 3D point cloud tracker, \ie, P2B, but also effectively deceive other unseen state-of-the-art approaches such as BAT and M^2Track, posing a new threat to 3D point cloud tracking.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2012

THE OBSERVATIONAL AND THEORETICAL TIDAL RADII OF GLOBULAR CLUSTERS IN M87

Globular clusters have linear sizes (tidal radii) which theory tells us are determined by their masses and by the gravitational potential of their host galaxy. To explore the relationship between observed and expected radii, we utilize the globular cluster population of the Virgo giant M87. Unusually deep, high signal-to-noise images of M87 are used to measure the effective and limiting radii of approximately 2000 globular clusters. To compare with these observations, we simulate a globular cluster population that has the same characteristics as the observed M87 cluster population. Placing these simulated clusters in the well-studied tidal field of M87, the orbit of each cluster is solved and the theoretical tidal radius of each cluster is determined. We compare the predicted relationship between cluster size and projected galactocentric distance to observations. We find that for an isotropic distribution of cluster velocities, theoretical tidal radii are approximately equal to observed limiting radii for R{sub gc} < 10 kpc. However, the isotropic simulation predicts a steep increase in cluster size at larger radii, which is not observed in large galaxies beyond the Milky Way. To minimize the discrepancy between theory and observations, we explore the effects of orbital anisotropy on cluster sizes, and suggest a possible orbital anisotropy profile for M87 which yields a better match between theory and observations. Finally, we suggest future studies which will establish a stronger link between theoretical tidal radii and observed radii.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Observational window effects on multi-object reverberation mapping

ABSTRACT Contemporary reverberation mapping campaigns are employing wide-area photometric data and high-multiplex spectroscopy to efficiently monitor hundreds of active galactic nuclei (AGNs). However, the interaction of the window function(s) imposed by the observation cadence with the reverberation lag and AGN variability time-scales (intrinsic to each source over a range of luminosities) impact our ability to recover these fundamental physical properties. Time dilation effects due to the sample source redshift distribution introduce added complexity. We present comprehensive analysis of the implications of observational cadence, seasonal gaps, and campaign baseline duration (i.e. the survey window function) for reverberation lag recovery. We find that the presence of a significant seasonal gap dominates the efficacy of any given campaign strategy for lag recovery across the parameter space, particularly for those sources with observed-frame lags above 100 d. Using the Australian Dark Energy Survey as a baseline, we consider the implications of this analysis for the 4MOST/Time-Domain Extragalactic Survey campaign providing concurrent follow-up of the Legacy Survey of Space and Time deep-drilling fields, as well as upcoming programmes. We conclude that the success of such surveys will be critically limited by the seasonal visibility of some potential field choices, but show significant improvement from extending the baseline. Optimizing the sample selection to fit the window function will improve survey efficacy.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

New Approach Methodology for Assessing Inhalation Risks of a Contact Respiratory Cytotoxicant: Computational Fluid Dynamics-Based Aerosol Dosimetry Modeling for Cross-Species and In Vitro Comparisons

Regulatory agencies are considering alternative approaches to assessing inhalation toxicity that utilizes in vitro studies with human cells and in silico modeling in lieu of additional animal studies. In support of this goal, computational fluid-particle dynamics models were developed to estimate site-specific deposition of inhaled aerosols containing the fungicide, chlorothalonil, in the rat and human for comparisons to prior rat inhalation studies and new human in vitro studies. Under bioassay conditions, the deposition was predicted to be greatest at the front of the rat nose followed by the anterior transitional epithelium and larynx corresponding to regions most sensitive to local contact irritation and cytotoxicity. For humans, simulations of aerosol deposition covering potential occupational or residential exposures (1–50 µm diameter) were conducted using nasal and oral breathing. Aerosols in the 1–5 µm range readily penetrated the deep region of the human lung following both oral and nasal breathing. Under actual use conditions (aerosol formulations >10 µm), the majority of deposited doses were in the upper conducting airways. Beyond the nose or mouth, the greatest deposition in the pharynx, larynx, trachea, and bronchi was predicted for aerosols in the 10–20 µm size range. Only small amounts of aerosols >20 µm penetrated past the pharyngeal region. Using the ICRP clearance model, local retained tissue dose metrics including maximal concentrations and areas under the curve were calculated for each airway region following repeated occupational exposures. These results are directly comparable with benchmark doses from in vitro toxicity studies in human cells leading to estimated human equivalent concentrations that reduce the reliance on animals for risk assessments.

63 RADIATION, THERMAL, AND OTHER ENVIRON. POLLUTAN↗
DOE OSTI2024

Tree-based algorithms for weakly supervised anomaly detection

Weakly supervised methods have emerged as a powerful tool for model-agnostic anomaly detection at the Large Hadron Collider (LHC). While these methods have shown remarkable performance on specific signatures such as dijet resonances, their application in a more model-agnostic manner requires dealing with a larger number of potentially noisy input features. In this paper, we show that using boosted decision trees as classifiers in weakly supervised anomaly detection gives superior performance compared to deep neural networks. Boosted decision trees are well known for their effectiveness in tabular data analysis. Our results show that they not only offer significantly faster training and evaluation times, but they are also robust to a large number of noisy input features. By using advanced gradient boosted decision trees in combination with ensembling techniques and an extended set of features, we significantly improve the performance of weakly supervised methods for anomaly detection at the LHC. This advance is a crucial step toward a more model-agnostic search for new physics. Published by the American Physical Society 2024

Astronomy & Astrophysics↗