Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “AI efficiency”

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 559 records · Page 31
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Public Reference Data for Megawatt-Scale Hydrogen Electrolysis - NLR Historical Wind

The U.S. Department of Energy and the National Laboratory of the Rockies (NLR) demonstrate hydrogen electrolysis from variable sources, hydrogen compression and storage, and hydrogen fuel cell power production using megawatt-scale equipment at NLR’s Flatirons Campus as part of the Advanced Research on Integrated Energy Systems (ARIES) initiative. This dataset represents part of that effort and is intended for academic, national laboratory, industrial, and other stakeholders to plan, design, and validate models of megawatt-scale hydrogen technologies and diverse energy infrastructure nationwide. These data provide a baseline for how existing hydrogen electrolysis technologies perform when coupled with various energy technologies. Future datasets will demonstrate how existing hydrogen fuel cell technologies can provide controllable, dispatchable, and variable power output for artificial intelligence (AI) data centers and other variable loads. This dataset entry describes hydrogen production by conducting a statistical analysis of historical wind data over a five-year period (2020-2025) from a single 1.5MW turbine manufactured by General Electric (GE) located at NLR’s Flatirons Campus, to generate an experimental test profile that was deployed on a 1.25-MW proton exchange membrane type MC250 electrolyzer system manufactured by Nel Hydrogen . [1] While the electrolyzer balance-of-plant supports up to 2.5 MW of electrolysis, NLR only has a single 1.25-MW electrolysis stack. The historical wind data provided several metrics, however, the analysis particularly focused on the measured power output by the wind turbine. The power output time series of data for each day was categorized by total energy generation and standard deviation, and the day that represented the highest combination of these two metrics was chosen – December 25th, 2022. This process was then repeated for a moving four-hour window within this day to identify the most statistically variable period. Finally, this four-hour period was scaled by 65% to match the 1.25 MW electrolyzer. The electrolysis system controls hydrogen production by varying DC current applied to the stack, from a maximum of 3000 A to a minimum safe operation of 300 A, or 10%. Because the current – voltage characteristic changes as the stack ages and efficiency degrades, the actual minimum safe operating power changes over time. The historical wind profiles were translated from power (kilowatts) to current (amperes) using a curve fit with calibration data and sent to the electrolyzer power supply at 1 Hz frequency. For more details on the statistical analysis process, see the presentation labeled “ Public Reference Data for Megawatt-Scale Hydrogen Electrolysis” provided with each data entry. These datasets report relevant hydrogen balance-of-plant and system data, all captured at 1 Hz, including hydrogen mass production measured with an Emerson Coriolis flow meter. Each .zip file represents a single wind turbine electrolysis experiment and is formatted as follows: {technology}_{scaling factor}-{electrolyzer ramp rate in amperes/second} For instance, “wind-GE1.5MW_0.65-400.zip” represents the hour-long experiment using historical data from the wind-GE1.5MW turbine, scaled to 65%, with the electrolyzer power supply set to a maximum ramp rate (gain and slew) of 400 A/s. Each .zip folder contains the following files: A .csv file containing raw data An .xlsx file explaining all the fields in the raw data. A .png plot showing the time series of hydrogen production, electrolysis power consumption, and wind power input. A PDF file detailing the historical wind data statistical analysis used to generate the wind profile. An experiment labeled “characterization_200.zip” demonstrates the MC250 electrolyzer steady-state response with 30-minute load steps for a total duration of 5 hours. Finally, a .csv file is provided with all simulated wind experiments combined into one dataset labeled "combined_historical_wind_experiments.csv". NLR also built an AI/machine-learning predictive model based on these datasets. The model ingests the electrolyzer current command in amperes, as well as various pressures and temperatures across the system, and predicts hydrogen output in kilograms per hour. The complete model can be found at https://huggingface.co/NatLabRockies/ptmelt-hydrogen-electrolysis [1] nelhydrogen.com/product/mc-series-electrolyser .

08 HYDROGEN↗
DOE OSTI2024

Stochastic machine learning via sigma profiles to build a digital chemical space

This work establishes a different paradigm on digital molecular spaces and their efficient navigation by exploiting sigma profiles. To do so, the remarkable capability of Gaussian processes (GPs), a type of stochastic machine learning model, to correlate and predict physicochemical properties from sigma profiles is demonstrated, outperforming state-of-the-art neural networks previously published. The amount of chemical information encoded in sigma profiles eases the learning burden of machine learning models, permitting the training of GPs on small datasets which, due to their negligible computational cost and ease of implementation, are ideal models to be combined with optimization tools such as gradient search or Bayesian optimization (BO). Gradient search is used to efficiently navigate the sigma profile digital space, quickly converging to local extrema of target physicochemical properties. While this requires the availability of pretrained GP models on existing datasets, such limitations are eliminated with the implementation of BO, which can find global extrema with a limited number of iterations. A remarkable example of this is that of BO toward boiling temperature optimization. Holding no knowledge of chemistry except for the sigma profile and boiling temperature of carbon monoxide (the worst possible initial guess), BO finds the global maximum of the available boiling temperature dataset (over 1,000 molecules encompassing more than 40 families of organic and inorganic compounds) in just 15 iterations (i.e., 15 property measurements), cementing sigma profiles as a powerful digital chemical space for molecular optimization and discovery, particularly when little to no experimental data is initially available.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Performance of the ATLAS muon triggers in Run 2

The performance of the ATLAS muon trigger system is evaluated with proton-proton ( pp ) and heavy-ion (HI) collision data collected in Run 2 during 2015–2018 at the Large Hadron Collider. It is primarily evaluated using events containing a pair of muons from the decay of Z bosons to cover the intermediate momentum range between 26 GeV and 100 GeV. Overall, the efficiency of the single-muon triggers is about 68% in the barrel region and 85% in the endcap region. The p T range for efficiency determination is extended by using muons from decays of J ψ mesons, W bosons, and top quarks. The performance in HI collision data is measured and shows good agreement with the results obtained in pp collisions. The muon trigger shows uniform and stable performance in good agreement with the prediction of a detailed simulation. Dedicated multi-muon triggers with kinematic selections provide the backbone to beauty, quarkonia, and low-mass physics studies. The design, evolution and performance of these triggers are discussed in detail.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

Characterization of the Cherenkov Photon Background for Low-noise Silicon Detectors in Space

Future space observatories that seek to perform imaging and spectroscopy of faint astronomical sources will require ultra-low-noise detectors that are sensitive over a broad wavelength range. Silicon charge-coupled devices (CCDs), such as EMCCDs, skipper CCDs, multi-amplifier sensing CCDs, and single-electron sensitive read out CCDs have demonstrated the ability to detect and measure single photons from X-ray energies to near the silicon band gap (∼1.1 μm), making them candidate technologies for this application. Here, in this context, we study a relatively unexplored source of low-energy background coming from Cherenkov radiation produced by energetic cosmic rays traversing a silicon detector. We present a model for Cherenkov photon production and absorption that is calibrated to laboratory data, and we use this model to characterize the residual background rate for ultra-low-noise silicon detectors in space. We study how the Cherenkov background rate depends on detector thickness, variations in solar activity, and the contribution of heavy cosmic ray species (Z > 2). We find that for thick silicon detectors, such as those required to achieve high quantum efficiency at long wavelengths, the rate of cosmic-ray-induced Cherenkov photon production is comparable to other detector and astrophysical backgrounds. We apply our Cherenkov background model to simulated spectroscopic observations of extra-solar planets, and we find that thick detectors continue to outperform their thinner counterparts at longer wavelengths despite a larger Cherenkov background rate. Furthermore, we find that minimal masking of cosmic-ray tracks continues to maximize the signal-to-noise ratio of very faint sources despite the existence of extended halos of Cherenkov photons.

Astronomical detectors↗
DOE OSTI2026

Ground and excited state gradients with end-to-end differentiable semiempirical quantum chemistry

Accurate and efficient gradients of molecular energy with respect to nuclear degrees of freedom are essential for geometry optimization and molecular dynamics, including simulations that go beyond the Born–Oppenheimer regime. A common approach involves deriving analytical formulas for new electronic structure methods, which is often conceptually difficult and requires tedious coding. Here, we implement analytical, semi-numerical, and automatic differentiation (AD)-based gradient pathways for semiempirical Hamiltonian models in the PYSEQM software package, leveraging both graphics processing unit (GPU) and central processing unit (CPU) architectures. We further extend these capabilities to excited states calculated using the configuration interaction singles and time-dependent Hartree–Fock ansätze. We benchmark wall time, peak memory usage, and accuracy across three molecular families of varying chemical complexity, including systems of up to a thousand atoms. For ground-state simulations, analytical and AD gradients achieve near-identical GPU runtimes, while semi-numerical gradients are slower on GPU but remain competitive on CPU. For excited states, both analytical and custom AD approaches using implicit differentiation show similar performance and low memory requirements, whereas gradients with full AD are memory-limited. AD gradients match analytical ones in accuracy across all tested systems, aided by a quaternion-based diatomic frame rotation for two-center quantities that ensures smooth energy surfaces. Overall, automatic differentiation emerges as a practical alternative to analytical gradients in semiempirical quantum chemistry, offering high accuracy while allowing seamless integration in AI-driven workflows and popular packages, such as PyTorch and JAX. Our results provide actionable guidance for selecting optimal gradient strategies in large-scale ground- and excited-state molecular dynamics simulations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Beam-induced backgrounds measured in the ATLAS detector during local gas injection into the LHC beam vacuum

Inelastic beam-gas collisions at the Large Hadron Collider (LHC), within a few hundred metres of the ATLAS experiment, are known to give the dominant contribution to beam backgrounds. These are monitored by ATLAS with a dedicated Beam Conditions Monitor (BCM) and with the rate of fake jets in the calorimeters. These two methods are complementary since the BCM probes backgrounds just around the beam pipe while fake jets are observed at radii of up to several metres. In order to quantify the correlation between the residual gas density in the LHC beam vacuum and the experimental backgrounds recorded by ATLAS, several dedicated tests were performed during LHC Run 2. Local pressure bumps, with a gas density several orders of magnitude higher than during normal operation, were introduced at different locations. The changes of beam-related backgrounds, seen in ATLAS, are correlated with the local pressure variation. In addition the rates of beam-gas events are estimated from the pressure measurements and pressure bump profiles obtained from calculations. Using these rates, the efficiency of the ATLAS beam background monitors to detect beam-gas events is derived as a function of distance from the interaction point. These efficiencies and characteristic distributions of fake jets from the beam backgrounds are found to be in good agreement with results of beam-gas simulations performed with theFluka Monte Carlo programme.

43 PARTICLE ACCELERATORS↗
DOE OSTI2026

Transforming jet flavour tagging at ATLAS

Jet flavour tagging enables the identification of jets originating from heavy-flavour quarks in proton–proton collisions at the Large Hadron Collider, playing a critical role in its physics programmes. This paper presents GN2, a transformer-based flavour tagging algorithm deployed by the ATLAS Collaboration that represents a different methodology compared to previous approaches. Designed to classify jets based on the flavour of their constituent particles, GN2 processes low-level tracking information in an end-to-end architecture and incorporates physics-informed auxiliary training objectives to enhance both interpretability and performance. Its performance is validated in both simulation and collision data. The measured c-jet (light-jet) rejection in data is improved by a factor of 3.5 (1.8) for a 70% b-jet tagging efficiency, compared to the previous algorithm. GN2 provides substantial benefits for physics analyses involving heavy-flavour jets, such as measurements of Higgs boson pair production and the couplings of bottom and charm quarks to the Higgs boson, and demonstrates the impact of advanced machine learning methods in experimental particle physics.

Characterization and analytical techniques↗
DOE OSTI2021

Performance of the ATLAS RPC detector and Level-1 muon barrel trigger at √(s)=13 TeV

The ATLAS experiment at the Large Hadron Collider (LHC) employs a trigger system consisting of a first-level hardware trigger (L1) and a software-based high-level trigger. The L1 muon trigger system selects muon candidates, assigns them to the correct LHC bunch crossing and classifies them into one of six transverse-momentum threshold classes. The L1 muon trigger system uses resistive-plate chambers (RPCs) to generate the muon-induced trigger signals in the central (barrel) region of the ATLAS detector. The ATLAS RPCs are arranged in six concentric layers and operate in a toroidal magnetic field with a bending power of 1.5 to 5.5 Tm. The RPC detector consists of about 3700 gas volumes with a total surface area of more than 4000 m2. This paper reports on the performance of the RPC detector and L1 muon barrel trigger using 60.8 fb-1 of proton-proton collision data recorded by the ATLAS experiment in 2018 at a centre-of-mass energy of 13 TeV. Detector and trigger performance are studied using Z boson decays into a muon pair. Measurements of the RPC detector response, efficiency, and time resolution are reported. Measurements of the L1 muon barrel trigger efficiencies and rates are presented, along with measurements of the properties of the selected sample of muon candidates. Measurements of the RPC currents, counting rates and mean avalanche charge are performed using zero-bias collisions. Finally, RPC detector response and efficiency are studied at different high voltage and front-end discriminator threshold settings in order to extrapolate detector response to the higher luminosity expected for the High Luminosity LHC.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

Configuration, Performance, and Commissioning of the ATLAS b-jet Triggers for the 2022 and 2023 LHC data-taking periods

In 2022 and 2023, the Large Hadron Collider produced approximately two billion hadronic interactions each second from bunches of protons that collide at a rate of 40 MHz. The ATLAS trigger system is used to reduce this rate to a few kHz for recording. Selections based on hadronic jets, their energy, and event topology reduce the rate to 𝒪(10) kHz while maintaining high efficiencies for important signatures resulting in b-quarks, but to reach the desired recording rate of hundreds of Hz, additional real-time selections based on the identification of jets containing b-hadrons (b-jets) are employed to achieve low thresholds on the jet transverse momentum at the High-Level Trigger. The configuration, commissioning, and performance of the real-time ATLAS b-jet identification algorithms for the early LHC Run 3 collision data are presented. These recent developments provide substantial gains in signal efficiency for critical signatures; for the Standard Model production of Higgs boson pairs, a 50% improvement in selection efficiency is observed in final states with four b-quarks or two b-quarks and two hadronically decaying τ-leptons.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

Knowledge Oriented Graph Unified Transformer (KOGUT) v0.1

KOGUT — Knowledge Oriented Graph Unified Transformer KOGUT implements the Relational Graph Transformer (RelGT) architecture for knowledge graph link prediction in biological domains, with a primary focus on microbial growth media prediction. While the original RelGT (arXiv:2505.10960) targets relational tables, time series, and multi-table databases, KOGUT adapts this architecture for heterogeneous biological knowledge graphs, providing first-in-class AI predictive models for microbial cultivation. Key Adaptations Beyond Original RelGT: - Knowledge Graph Focus: Applied to biological KGs with semantic node types (taxa, chemicals, media, phenotypes, environments) versus generic relational database tables, trained on the KG-Microbe knowledge graph (1.3M entities, 2.9M edges, 24 relation types). - Multimodal Node Encoding: Integrates node labels, categories, descriptions, and synonyms from KG metadata through learned embedding layers—adapting relational column features to graph node attributes with textual semantics. - Extended K-Hop Subgraph Strategy: Optimized neighborhood sampling (3-hop default, configurable up to 200 nodes) tuned for sparse biological networks, building on the original local-global attention framework with biological relation preservation. - Biolink Predicate Preservation: Type-specific transformations for 24 biological edge semantics (occurs_in, consumes, produces, has_phenotype, subclass_of) beyond standard relational foreign keys, enabling multi-relation link prediction. - Inductive Learning Support: Enables zero-shot predictions for novel taxa through feature-based embeddings (temperature, oxygen requirements, gram stain, cell shape), extending the original transductive relational benchmark scope to uncultured microorganisms. CheapSOTA Performance Optimizations (This Distribution): - VQ-EMA Centroid Attention: Vector quantization with exponential moving average for improved global context modeling (+5-10% MRR improvement). - HDF5 Precomputed Data Loading: One-time preprocessing of k-hop subgraphs to eliminate redundant graph traversals (2-5× training speedup). - Distributed Data Parallel Training: Multi-GPU support for scaling to larger knowledge graphs (tested on 4× NVIDIA A100 GPUs at NERSC Perlmutter). - Mixed Precision Training: Automatic mixed precision (AMP) for memory efficiency and faster training. Advantages Over Standard Knowledge Graph Embedding Models: Combines RelGT's proven multi-element tokenization (features, type, hop, structure) with graph-native biological representations, enabling interpretable link prediction across heterogeneous entities that standard embedding models (TransE, RotatE, ComplEx) and table-based transformers cannot directly model. Achieves near-perfect performance on microbial growth media prediction (MRR: 0.9966, Precision@1: 0.9932, Hit@10: 1.0000) while maintaining explainability through attention-based reasoning over biological pathways. Training Data: - KG-Microbe merged knowledge graph: 1,379,337 nodes, 2,960,472 edges - 24 biological relation types including taxonomic hierarchies, metabolic interactions, phenotype associations, and environmental relationships - Primary prediction task: Growth media suitability for microbial taxa (biolink:occurs_in, 50K edges) - Multi-relation capability: Predicts links for any of the 24 relation types, including chemical consumption/production, phenotype associations, and taxonomic classification Citation: Original RelGT Architecture: Dwivedi et al., "Relational Graph Transformer", arXiv:2505.10960, 2025 KOGUT Implementation: Knowledge Oriented Graph Unified Transformer for Microbial Growth Media Prediction Developed at Lawrence Berkeley National Laboratory (LBNL) Trained on NERSC Perlmutter supercomputer

Joachimiak, Marcin [Lawrence Berkeley National Lab↗
DOE OSTI2021

In Situ Formation of Polycyclic Aromatic Hydrocarbons as an Artificial Hybrid Layer for Lithium Metal Anodes

Nonuniform Li deposition causes dendrites and low Coulombic efficiency (CE), seriously hindering the practical applications of Li metal. Herein, we developed an artificial solid-state interphase (SEI) with planar polycyclic aromatic hydrocarbons (PAHs) on the surface of Li metal anodes by a facile in situ formation technology. The resultant dihydroxyviolanthron (DHV) layers serve as the protective layer to stabilize the SEI. In addition, the oxygen-containing functional groups in the soft and conformal SEI film can regulate the diffusion and transport of Li ions to homogenize the deposition of Li metal. The artificial SEI significantly improves the CEs and shows superior cyclability of over 1000 h at 4 mAh cm –2 . The LiFePO 4 /Li cell (2.8 mAh cm –2 ) enables a long cyclability for 300 cycles and high CEs of 99.8%. Furthermore, this work offers a new strategy to inhibit Li dendrite growth and enlightens the design on stable SEI for metal anodes.

25 ENERGY STORAGE↗
DOE OSTI2025

Transforming Energy Through Computational Excellence: A View From NREL

At the National Renewable Energy Laboratory (NREL)—a U.S. Department of Energy laboratory—computational science, high-performance computing, applied mathematics, advanced computer science, visualization, and data play a pivotal role in advancing energy abundance, affordability, security, and reliability. From fundamental scientifc discovery to systems engineering and analysis, NREL researchers tackle market-relevant challenges to develop solutions for an independent energy system that is reliable, resilient and secure. Collaborative partnerships with industry, government, and academia ensure that our research remains cutting edge, impactful, applicable, and aligned with real-world energy needs. This special issue of Computing in Science & Engineering highlights exemplary NREL projects where computational tools and methodologies drive discovery and accelerate innovation in scalable and integrated energy systems. The featured articles explore the role of computational modeling, high-performance computing, generative AI, and adaptive computing in advancing independent energy solutions, optimizing sustainability research, and enhancing decision-making for energy solutions using a broad mix of energy technologies. Here, these contributions demonstrate how NREL’s computational research bridges the gap between theoretical advancements and practical implementation, emphasizing interdisciplinary collaboration and a commitment to innovation, with a focus on translating computational excellence into real-world impact, thus accelerate progress toward national energy goals. By showcasing cutting-edge research at the intersection of computational science and energy systems, this issue aims to inspire and inform researchers, practitioners, and policymakers dedicated to shaping a more reliable energy future.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Hierarchical Multi-agent Large Language Model Reasoning for Autonomous Heterogeneous Catalyst Discovery

Artificial intelligence is reshaping scientific exploration, but most methods automate procedural tasks without engaging in scientific reasoning, limiting autonomy in discovery. We demonstrate that hierarchical agentic large language model reasoning can efficiently drive simulation and scientific exploration. Across two chemical applications, CO adsorption on Cu surface transition metal adatoms and on M–N–C catalysts, reasoning-guided exploration reduces required atomistic simulations by up to 90% relative to heuristic or random selection. Comparisons across single-agent, multi-agent, and stochastic baselines show that hierarchical strategies yield more coherent and information-efficient search trajectories. Reasoning traces reveal chemically grounded decisions that cannot be explained by semantic bias or stochastic sampling. We realize these agentic reasoning strategies in Materials Agents for Simulation and Theory in Electronic-structure Reasoning (MASTER), a multimodal system that translates natural language into density functional theory workflows. Altogether, multi-agent collaboration accelerates heterogeneous catalyst discovery and marks a step toward more autonomous, reasoning-guided scientific exploration.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2025

Chemical classification program synthesis using generative artificial intelligence

Accurately classifying chemical structures is essential for cheminformatics and bioinformatics, including tasks such as identifying bioactive compounds of interest, screening molecules for toxicity to humans, finding non-organic compounds with desirable material properties, or organizing large chemical libraries for drug discovery or environmental monitoring. However, manual classification is labor-intensive and difficult to scale to large chemical databases. Existing automated approaches either rely on manually constructed classification rules, or are deep learning methods that lack explainability. This work presents an approach that uses generative artificial intelligence to automatically write chemical classifier programs for classes in the Chemical Entities of Biological Interest (ChEBI) database. These programs can be used for efficient deterministic run-time classification of SMILES structures, with natural language explanations. The programs themselves constitute an explainable computable ontological model of chemical class nomenclature, which we call the ChEBI Chemical Class Program Ontology (C3PO). We validated our approach against the ChEBI database, and compared our results against deep learning models and a naive SMARTS pattern based classifier. C3PO outperforms the naive classifier, but does not reach the performance of state of the art deep learning methods. However, C3PO has a number of strengths that complement deep learning methods, including explainability and reduced data dependence. C3PO can be used alongside deep learning classifiers to provide an explanation of the classification, where both methods agree. The programs can be used as part of the ontology development process, and iteratively refined by expert human curators.

Artificial Intelligence↗
DOE OSTI2024

Calibration of a soft secondary vertex tagger using proton-proton collisions at s = 13 TeV with the ATLAS detector

Several processes studied by the ATLAS experiment at the Large Hadron Collider produce low-momentum b-flavored hadrons in the final state. This paper describes the calibration of a dedicated tagging algorithm that identifies b-flavored hadrons outside of hadronic jets by reconstructing the soft secondary vertices originating from their decays. The calibration is based on a proton-proton collision dataset at a center-of-mass energy of 13 TeV corresponding to an integrated luminosity of 140 fb -1 . Scale factors used to correct the algorithm’s performance in simulated events are extracted for the b-tagging efficiency and the mistag rate of the algorithm using a data sample enriched in $t\overline{t}$ events. Several orthogonal measurement regions are defined, binned as a function of the multiplicities of soft secondary vertices and jets containing a b-flavored hadron in the event. The mistag rate scale factors are estimated separately for events with low and high average numbers of interactions per bunch crossing. The results, which are derived from events with low missing transverse momentum, are successfully validated in a phase space characterized by high missing transverse momentum and therefore are applicable to new physics searches carried out in either phase space regime.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
NASA NTRS2019

Deep Learning Method for Detecting Precursors to Adverse Events

With the recent advancements in Deep Learning methods, the ability to model large complex heterogeneous data sets are fundamentally changing industry and research. Coupled with hardware improvements, and ease of implementation, a wide variety of deep neural network architectures can quickly be developed to solve a sweeping range of problems such as: object detection in images, automatic healthcare diagnosis using heterogenous data sources, real time language translating and sentence prediction, upscaling low resolution images, and forecasting of multivariate timeseries. Generally, many of these architectures outperform classical machine learning approaches in their respective tasks, however, this typically comes at a cost of interpretability. These black box algorithms generally suffer from lack of transparency in both model complexity as well as the rationale behind the prediction. This lack of comprehension, is driving an emerging area of interest in “Explainable AI”. An algorithm called: “Deep Temporal Multiple Instance Learning”1 was a recently developed to identify precursors to adverse events and has been applied in the aviation domain. The deep learning architecture is designed to capture the evolution of the probability of the outcome over the time preceding the adverse event using a multiple instance learning approach as illustrated in Figure 1. Precursors are defined when the probability of the event has exceeded a threshold at some point in the timeseries, at which point, a sensitivity analysis is performed to determine contributing factors. The contributing factors are used to explain and define the precursor during the periods where the probability score is high. The identified contributing factors are then presented to subject matter experts to provide objective insights into the leading factors associated with the particular adverse event. The algorithm has been tested on flight data from a commercial airline and has the ability to discover precursors to known adverse events that take the form of safety critical operations, such as unstable approach events on final approach. Apart from detecting precursors to adverse events, the converse can also be leveraged to discover corrective actions. These positive actions manifest themselves as periods in the timeseries when the precursor score has been lowered from an elevated state; meaning that if the system had been left uncorrected, it would have eventually reached the adverse event state. Characterizing these state changes can help identify successful interventions that may not have been known before. Policy makers and procedure designers can use this additional knowledge to craft more safety and efficient resilient procedures for future operations and therefore improve the overall performance of the National Airspace.

Matthews, Bryan L.↗
DOE OSTI2025

Simulating Atmospheric Processes in Earth System Models and Quantifying Uncertainties With Deep Learning Multi‐Member and Stochastic Parameterizations

Abstract Deep learning is a powerful tool to represent subgrid processes in climate models, but many application cases have so far used idealized settings and deterministic approaches. Here, we develop stochastic parameterizations with calibrated uncertainty quantification to learn subgrid convective and turbulent processes and surface radiative fluxes of a superparameterization embedded in an Earth System Model (ESM). We explore three methods to construct stochastic parameterizations: (a) a single Deep Neural Network (DNN) with Monte Carlo Dropout; (b) a multi‐member parameterization; and (c) a Variational Encoder Decoder with latent space perturbation. We show that the multi‐member parameterization improves the representation of convective processes, especially in the planetary boundary layer, compared to individual DNNs. The respective uncertainty quantification illustrates that methods (b) and (c) are advantageous compared to a dropout‐based DNN parameterization regarding the spread of convective processes. Hybrid simulations with our best‐performing multi‐member parameterizations remained challenging and crash within the first days. Therefore, we develop a pragmatic partial coupling strategy relying on the superparameterization for condensate emulation. Partial coupling reduces the computational efficiency of hybrid Earth‐like simulations but enables model stability over 5 months with our multi‐member parameterizations. However, our hybrid simulations exhibit biases in thermodynamic fields and differences in precipitation patterns. Despite this, the multi‐member parameterizations enable improvements in reproducing tropical extreme precipitation compared to a traditional convection parameterization. Despite these challenges, our results indicate the potential of a new generation of multi‐member machine learning parameterizations leveraging uncertainty quantification to improve the representation of stochasticity of subgrid effects.

Behrens, Gunnar [Deutsches Zentrum für Luft‐ und R↗
DOE OSTI2023

Calibration of the light-flavour jet mistagging efficiency of the b -tagging algorithms with Z+jets events using 139 $\textrm{fb}^{-1}$ of ATLAS proton–proton collision data at $\sqrt{s} = 13$ TeV

The identification of b -jets, referred to as b -tagging, is an important part of many physics analyses in the ATLAS experiment at the Large Hadron Collider and an accurate calibration of its performance is essential for high-quality physics results. This publication describes the calibration of the light-flavour jet mistagging efficiency in a data sample of proton–proton collision events at $\sqrt{s} = 13$ TeV corresponding to an integrated luminosity of 139 fb -1 . The calibration is performed in a sample of Z bosons produced in association with jets. Due to the low mistagging efficiency for light-flavour jets, a method which uses modified versions of the b-tagging algorithms referred to as flip taggers is used in this work. A fit to the jet-flavour-sensitive secondary-vertex mass is performed to extract a scale factor from data, to correct the light-flavour jet mistagging efficiency in Monte Carlo simulations, while simultaneously correcting the b -jet efficiency. With this procedure, uncertainties coming from the modeling of jets from heavy-flavour hadrons are considerably lower than in previous calibrations of the mistagging scale factors, where they were dominant. The scale factors obtained in this calibration are consistent with unity within uncertainties.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗