SEARCH · Search NASA

Results for “Large Sets”

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 55 records · Page 3

All sources NASA NTRS DOE OSTI

DOE OSTI2022

237 Np Mössbauer Isomer Shifts: A Lesson About the Balance of Static and Dynamic Electron Correlation in Heavy Element Complexes

A large set of neptunium compounds with different oxidation states (III to VII) was assembled to study the Mössbauer isomer shift by wave function calculations and better understand covalency in f-elements complexes. The contact density approach was used to calculate the isomer shift using complete active space self-consistent field (CASSCF) multiconfiguration wave functions, as well as matrix product states [from Density Matrix Renormalization Group (DMRG) algorithms] for large active spaces. Dynamic correlation effects for the isomer shifts were treated via CASPT2 energy derivatives with respect to the nuclear radius. The CASSCF calculations appear to produce different orbital overlocalization errors for low and high Np oxidation states. For compounds with low Np oxidation numbers, the errors can be attributed to the overlocalization of the 5f orbitals. For the compounds with high Np oxidation numbers, the main errors arise from an overlocalization of ligand orbitals and concomitant to weak donation bonding. Attempts to mitigate the overlocalization errors with large active spaces using DMRG were only partially successful, showing that explicit treatment of dynamic correlation is necessary for accurate predictions of Mössbauer isomer shifts. The CASPT2 calculations perform very satisfactorily. For a subset of Np compounds, both static and dynamic correlation effects were substantial. Furthermore, a rational active space selection based on orbital entanglement diagrams proved beneficial for determining the optimal reference wave function.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Representations and strategies for transferable machine learning improve model performance in chemical discovery

Strategies for machine-learning (ML)-accelerated discovery that are general across material composition spaces are essential, but demonstrations of ML have been primarily limited to narrow composition variations. By addressing the scarcity of data in promising regions of chemical space for challenging targets such as open-shell transition-metal complexes, general representations and transferable ML models that leverage known relationships in existing data will accelerate discovery. Over a large set (~1000) of isovalent transition-metal complexes, we quantify evident relationships for different properties (i.e., spin-splitting and ligand dissociation) between rows of the Periodic Table (i.e., 3d/4d metals and 2p/3p ligands). We demonstrate an extension to the graph-based revised autocorrelation (RAC) representation (i.e., eRAC) that incorporates the group number alongside the nuclear charge heuristic that otherwise overestimates dissimilarity of isovalent complexes. To address the common challenge of discovery in a new space where data are limited, we introduce a transfer learning approach in which we seed models trained on a large amount of data from one row of the Periodic Table with a small number of data points from the additional row. We demonstrate the synergistic value of the eRACs alongside this transfer learning strategy to consistently improve model performance. Analysis of these models highlights how the approach succeeds by reordering the distances between complexes to be more consistent with the Periodic Table, a property we expect to be broadly useful for other material domains.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2026

Subseasonal Forecasting and MJO Teleconnections in Machine Learning Weather Prediction Models

Abstract In recent years, machine‐learning (ML) models trained on reanalysis data have rivaled physics‐based forecast models in terms of performance skill for global weather forecasting. With increased rollout stability, the question of how these models perform for subseasonal to seasonal (S2S, week 3–8) forecasting has emerged. In this study we run a large set of subseasonal hindcasts over 2004–2023 to evaluate two ML weather forecast models at the S2S time scale, SFNO‐HENS (Nvidia, fully ML) and NeuralGCM (Google Research, hybrid). Corresponding hindcasts from the European Centre for Medium‐Range Weather Forecasts (ECMWF) are used as a baseline for comparison to a physics‐based model. Because our focus is on predicting moisture transport over the Western United States between October and March, we evaluate the models' prediction skill for the Madden‐Julian Oscillation (MJO) and its associated teleconnections in the North Pacific. We find that both ML models are competitive with the ECWMF model, with comparable skill in predicting the North Pacific large‐scale circulation and the MJO at week 3 and beyond. Even though overall the mid‐latitude subseasonal prediction skill remains low, the ML models exhibit interesting behavior such as a realistic propagation of the MJO across the Maritime Continent and realistic teleconnections. A SFNO‐HENS sensitivity experiment with altered initial conditions in the tropics demonstrates the stability of the model, and it illustrates the capability of ML models to represent important physical processes of the atmosphere at the S2S time scale. Plain Language Summary Predicting weather patterns and precipitation a few weeks in advance (subseasonal time scale) is of great interest for stakeholders such as water managers in the Southwest United States (US), where arid conditions prevail. Subseasonal forecasts from traditional weather forecast models exhibit low skill in the region, limiting their applicability. Here we examine whether the recent breakthrough in weather forecasting made with machine learning/artificial intelligence models can translate to improved subseasonal forecasts. Recently‐developed machine learning models exhibit comparable skill to a state‐of‐the‐art physics‐based model for predicting weather patterns in the North Pacific/North America region, and associated moisture transport. The same applies to their skill in predicting the tropical pattern, the Madden‐Julian Oscillation, and its important remote perturbations over the midlatitude East Pacific and Southwest US. Additionally, a perturbation experiment carried out with one of the machine learning models illustrates their ability to not only predict the evolution of atmospheric fields, but also to learn and represent physical processes such as tropics‐extratropics Rossby wave propagation. Key Points Two machine learning weather forecast models exhibit state‐of‐the‐art prediction skill at the subseasonal time scale in the Pacific sector The models equal ECWMF in terms of Madden‐Julian oscillation (MJO) prediction skill, and they accurately predict the MJO propagation and associated teleconnections The two machine‐learning models represent key physical processes for subseasonal prediction, despite being trained for weather forecasting

Peings, Yannick↗

DOE OSTI2024

Direct pulse-level compilation of arbitrary quantum logic gates on superconducting qutrits

Advanced simulations and calculations on quantum computers require high-fidelity implementations of quantum operations. The universal gateset approach builds complex unitaries from a small set of primitive gates, often resulting in a long gate sequence, which is typically a leading factor in the total accumulated error. Compiling a complex unitary for processors with higher-dimensional logical elements, such as qutrits, exacerbates the accumulated error per unitary, since an even longer gate sequence is required. Optimal control methods promise time- and resource-efficient compact gate sequences and, therefore, higher fidelity. These methods generate pulses that can directly implement any complex unitary on a quantum device. In this work, we demonstrate that any arbitrary qubit and qutrit gate can be realized with high fidelity, which can significantly reduce the length of a gate sequence. We generate and test pulses for a large set of randomly selected arbitrary unitaries on several quantum processing units (QPUs): the Lawrence Livermore National Laboratory Quantum Device and Integration Testbed’s (QuDIT’s) standard QPU and three of Rigetti’s QPUs: Ankaa-2, Ankaa-9Q-1, and Aspen-M-3. On the QuDIT platform’s standard QPU, the average fidelity of random qutrit gates is 97.9 ± 0.5% measured with conventional QPT and 98.8 ± 0.6% from QPT with gate folding. Rigetti’s Ankaa-2 achieves random qubit gates with an average fidelity of 98.4 ± 0.5% (conventional QPT) and 99.7 ± 0.1% (QPT with gate folding). On Ankaa-9Q-1 and Aspen-M-3, the average fidelities with conventional qubit QPT measurements were higher than 99% (see Appendix). Here we show that optimal control gates are robust to drift for at least 3 h and that the same calibration parameters can be used for all implemented gates. Our work promises that the calibration overheads for optimal control gates can be made small enough to enable efficient quantum circuits based on this technique.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2024

Generating Euler Diagrams Through Combinatorial Optimization

Abstract Can a given set system be drawn as an Euler diagram? We present the first method that correctly decides this question for arbitrary set systems if the Euler diagram is required to represent each set with a single connected region. If the answer is yes, our method constructs an Euler diagram. If the answer is no, our method yields an Euler diagram for a simplified version of the set system, where a minimum number of set elements have been removed. Further, we integrate known wellformedness criteria for Euler diagrams as additional optimization objectives into our method. Our focus lies on the computation of a planar graph that is embedded in the plane to serve as the dual graph of the Euler diagram. Since even a basic version of this problem is known to be NP‐hard, we choose an approach based on integer linear programming (ILP), which allows us to compute optimal solutions with existing mathematical solvers. For this, we draw upon previous research on computing planar supports of hypergraphs and adapt existing ILP building blocks for contiguity‐constrained spatial unit allocation and the maximum planar subgraph problem. To generate Euler diagrams for large set systems, for which the proposed simplification through element removal becomes indispensable, we also present an efficient heuristic. We report on experiments with data from MovieDB and Twitter. Over all examples, including 850 non‐trivial instances, our exact optimization method failed only for one set system to find a solution without removing a set element. However, with the removal of only a few set elements, the Euler diagrams can be substantially improved with respect to our wellformedness criteria.

Computer Science↗

DOE OSTI2023

Photometric metallicity for 694 233 Galactic giant stars from Gaia DR3 synthetic Strömgren photometry: Metallicity distribution functions of halo substructures

We used two previous calibrations and the standardised synthetic photometry in the v, b, and y Strömgren passbands from Gaia DR3 BP/RP spectra to obtain photometric metallicities for a selected sample of 694 233 old Galactic giant stars having |b|> 20.0° and parallax uncertainties lower than 10%. The zero point of both sets of photometric metallicities has been shifted to ensure an optimal match with the spectroscopic [Fe/H] values for 44 785 stars in common with APOGEE DR17, focusing on the metallicity range where they provide the highest accuracy. The metallicities derived in this way from one calibration display a typical accuracy of ≲0.1 dex and 1σ precision ≲0.2 dex in the range -2.2 ≲ [Fe/H] ≲ - 0.4, while they show a systematic trend with [Fe/H] at higher metallicity, beyond the applicability range of the relation. Those derived from the other calibration display, in general, reduced precision, and lower accuracy in the metal-poor regime, but have a median accuracy < 0.05 dex for [Fe/H] ≥ - 0.8. These results are confirmed and, consequently, the metallicities are validated, by comparison with large sets of spectroscopic metallicities from various surveys. The newly obtained metallicities are used to derive metallicity distributions for several previously identified substructures in the Galactic halo with an unprecedented number of stars. The catalogue including the two sets of metallicities and the associated uncertainties is made publicly available.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2020

Sensitivity of surface temperature to oceanic forcing via q-Flux Green’s function experiments. Part III: asymmetric response to warming and cooling

Climate response is often assumed to be linear in climate sensitivity studies. However, by examining the surface temperature (TS) response to pairs of oceanic forcings of equal amplitude but opposite sign in a large set of local q-flux perturbation experiments with CAM5 coupled to a slab, we find strong asymmetry in TS responses to the heating and cooling forcings, indicating a strong nonlinearity intrinsic to the climate system examined. Regardless of where the symmetric forcing is placed, the cooling response to the negative forcing always exceeds the warming to the positive forcing, implying an intrinsic inclination towards cooling of our current climate. Thus, the ongoing global warming induced by increasing greenhouse gases may have already been alleviated by the asymmetric component of the response. The common asymmetry in TS response peaks in high latitudes, especially along sea ice edges, with notable seasonal dependence. Decomposition into different radiative feedbacks through a radiative kernel indicates that the asymmetry in the TS response is realized largely through lapse rate and albedo feedbacks. Further process interference experiments disabling the seasonal cycle and/or sea ice reveal that the asymmetry originates ultimately from the presence of the sea ice component and is further amplified by the seasonal cycle. The fact that a pair of opposite tropical q-flux forcings can excite very similar asymmetric response as a pair placed at 55°S strongly suggests the asymmetric response is a manifestation of an internal mode of the climate model system.

Liu, Fukai↗

DOE OSTI2024

Estimating the Contributions to Human Error Probability from the Convolution of the Distribution of Time Available and Time Required

As part of their duties, Human Reliability Analysis must often evaluate if crews in nuclear power plants (NPPs) can complete tasks associated with a human-failure event within time limits. For example, the time required in NPP scenarios is determined by systematic and structured walkthroughs, feasibility studies, recorded times from training exercises, and interviews with experienced operators and experts. Typically, a point estimate is derived for the estimate (mean, maximum, or 95th percentile of time required). Using point-estimate values can mask the risk associated with variability among crews, plant conditions and set-up, environmental conditions, and other impact factors under which these actions are executed. While point estimates for time required and time available have served the industry well, without considering the uncertainty they could lead to biased understanding about the risk. The Integrated Human Event Analysis System - General Methodology (IDHEAS-G) model (developed by the US Nuclear Regulatory Commission, NRC) for human error probability calculates human error probability by summing two probabilities: insufficient time and cognitive error. As such, the model takes a more holistic approach by considering the full distributions for time required and time available to calculate the human error probability because the time available to complete the task is insufficient. In this study, we expand on the work of the NRC and discuss methods for estimating these time considerations. For example, for the time required, the impact of Performance Influencing Factors (PIFs) on the distribution was divided into impacts that are aleatory in nature, such as crew-to-crew variability, and those that are epistemic (i.e., the PIFs). Starting with the factors that introduce aleatory uncertainty, a first-order distribution was developed from a large set of time required (i.e., NPP task completion times) data for the range of operator actions that occur in the NPP control room under simulated accident conditions. The first-order distribution can then be adjusted to account for epistemic uncertainty using research associated with the impact of applicable PIFs on the time required. We also develop guidance for analysts to address the probability distributions for the time available. The guidance we developed on how to estimate time required and time available distributions is based on the identification of pertinent research and data, data analyses, and expert knowledge elicitation.

human error probability, human performance, time e↗

DOE OSTI2025

Environmental Conditions Affecting Global Mesoscale Convective System Occurrence

Abstract The ERA5 environments of mesoscale convective systems (MCSs), tracked from satellite observations, are assessed over a 20-yr period. The use of a large set of MCS tracks allows us to robustly test the sensitivity of the results to factors such as region, latitude, and diurnal cycle. We aim to provide novel information on environments of observed MCSs for assessments of global atmospheric models and to improve their ability to simulate MCSs. Statistical analysis of all tracked MCSs is performed in two complementary ways. First, we investigate the environments when an MCS has occurred at different spatial scales before and after MCS formation. Several environmental variables are found to show marked changes before MCS initiation, particularly over land. The vertically integrated moisture flux convergence shows a robust signal across different regions and when considering MCS initiation diurnal cycle. We also found spatial scale dependence of the environments between 200 and 500 km, providing new evidence of a natural length scale for use with MCS parameterization. In the second analysis, the likelihood of MCS occurrence for given environmental conditions is evaluated, by considering all environments and determining the probability of being in an MCS core or shield region. These are compared to analogous non-MCS environments, allowing discrimination between conditions suitable for MCS and non-MCS occurrence. Three environmental variables are found to be useful predictors of MCS occurrence: total column water vapor, midlevel relative humidity, and total column moisture flux convergence. Such relations could be used as trigger conditions for the parameterization of MCSs, thereby strengthening the dependence of the MCS scheme on the environment. Significance Statement Large storm systems called mesoscale convective systems form across Earth. These are collections of thunderstorms, with associated high-level clouds that produce substantial, lighter rainfall and modulate Earth’s energy balance. They produce hazardous weather conditions, such as floods and high winds, and are responsible for a high percentage of rainfall in many regions globally. We investigate the environmental conditions under which they form, so that we can understand the spatial extent of the environment which is important for their formation, and also where and when the effects of these storms might be felt. The novel information generated here should help improve the representation of these storms in weather and climate models, improving the prediction of rainfall, thunderclouds, and high-level clouds.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2022

Origin of Plutonium-244 in the Early Solar System

We investigate the origin in the early Solar System of the short-lived radionuclide 244 Pu (with a half life of 80 Myr) produced by the rapid (r) neutron-capture process. We consider two large sets of r-process nucleosynthesis models and analyse if the origin of 244 Pu in the ESS is consistent with that of the other r and slow (s) neutron-capture process radioactive nuclei. Uncertainties on the r-process models come from both the nuclear physics input and the astrophysical site. The former strongly affects the ratios of isotopes of close mass ( 129 I/ 127 I, 244 Pu/ 238 U, and 247 Pu/ 235 U). The 129 I/ 247 Cm ratio, instead, which involves isotopes of a very different mass, is much more variable than those listed above and is more affected by the physics of the astrophysical site. We consider possible scenarios for the evolution of the abundances of these radioactive nuclei in the galactic interstellar medium and verify under which scenarios and conditions solutions can be found for the origin of 244 Pu that are consistent with the origin of the other isotopes. Solutions are generally found for all the possible different regimes controlled by the interval (δ) between additions from the source to the parcel of interstellar medium gas that ended up in the Solar System, relative to decay timescales. If r-process ejecta in interstellar medium are mixed within a relatively small area (leading to a long δ), we derive that the last event that explains the 129 I and 247 Cm abundances in the early Solar System can also account for the abundance of 244 Pu. Due to its longer half life, however, 244 Pu may have originated from a few events instead of one only. If r-process ejecta in interstellar medium are mixed within a relatively large area (leading to a short δ), we derive that the time elapsed from the formation of the molecular cloud to the formation of the Sun was 9-16 Myr.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2023

Universal Polarization Transformations: Spatial Programming of Polarization Scattering Matrices Using a Deep Learning‐Designed Diffractive Polarization Transformer

Abstract Controlled synthesis of optical fields having nonuniform polarization distributions presents a challenging task. Here, a universal polarization transformer is demonstrated that can synthesize a large set of arbitrarily‐selected, complex‐valued polarization scattering matrices between the polarization states at different positions within its input and output field‐of‐views (FOVs). This framework comprises 2D arrays of linear polarizers positioned between isotropic diffractive layers, each containing tens of thousands of diffractive features with optimizable transmission coefficients. After its deep learning‐based training, this diffractive polarization transformer can successfully implement N i N o = 10 000 different spatially‐encoded polarization scattering matrices with negligible error, where N i and N o represent the number of pixels in the input and output FOVs, respectively. This universal polarization transformation framework is experimentally validated in the terahertz spectrum by fabricating wire‐grid polarizers and integrating them with 3D‐printed diffractive layers to form a physical polarization transformer. Through this set‐up, an all‐optical polarization permutation operation of spatially‐varying polarization fields is demonstrated, and distinct spatially‐encoded polarization scattering matrices are simultaneously implemented between the input and output FOVs of a compact diffractive processor. This framework opens up new avenues for developing novel devices for universal polarization control and may find applications in, e.g., remote sensing, medical imaging, security, material inspection, and machine vision.

Optical neural networks↗

DOE OSTI2026

Engineering of 2‐ketoacid Decarboxylases for Production of Isobutanol and Other Fusel Alcohols in Saccharomyces cerevisiae

Isobutanol is a fusel alcohol that can be produced microbially for use as a biofuel or upgraded into sustainable aviation fuel (SAF). A key enzyme in the isobutanol biosynthetic pathway is 2-ketoacid decarboxylase (KDC), which irreversibly decarboxylates 2-ketoisovalerate (KIV) to yield isobutyraldehyde. However, many previously characterized KDC enzymes also act promiscuously on other 2-ketoacids, (e.g., pyruvate) to produce a related aldehyde (e.g., acetaldehyde). This unwanted side reaction is especially important when isobutanol is produced in Saccharomyces cerevisiae (S. cerevisiae) because it leads to pyruvate being diverted to ethanol. In order to make S. cerevisiae a strict isobutanologen, a KDC enzyme that is specific for KIV must be deployed. In this study, we used a combination of cell-based and in vitro enzyme assays to investigate KDC substrate specificity, characterizing a large set of homologs for KIV, pyruvate, and phenylpyruvate (PPV) activity. A diverse range of substrate specificities was discovered, and some previously uncharacterized KDCs were revealed to have high KIV activity and low pyruvate activity. Multi-site saturation mutagenesis (SSM) of one of these KDCs identified mutants with increased KIV activity, while maintaining low levels of pyruvate activity. In a KIV bioconversion experiment, bioprospected and engineered KDCs allowed similar KIV consumption to when using the previously characterized Lactococcus lactis KdcA, though with some ethanol also produced. The KDCs identified here show promise for production of isobutanol and other alcohols derived from 2-ketoacids, and the dataset of newly characterized KDCs can inform future efforts to understand and engineer substrate specificity in KDCs.

2-ketoacid decarboxylase↗

DOE OSTI2024

Enhancing 2‐Pyrone Synthase Efficiency by High‐Throughput Mass‐Spectrometric Quantification and In Vitro/In Vivo Catalytic Performance Correlation

Abstract Engineering efficient biocatalysts is essential for metabolic engineering to produce valuable bioproducts from renewable resources. However, due to the complexity of cellular metabolic networks, it is challenging to translate success in vitro into high performance in cells. To meet such a challenge, an accurate and efficient quantification method is necessary to screen a large set of mutants from complex cell culture and a careful correlation between the catalysis parameters in vitro and performance in cells is required. In this study, we employed a mass‐spectrometry based high‐throughput quantitative method to screen new mutants of 2‐pyrone synthase (2PS) for triacetic acid lactone (TAL) biosynthesis through directed evolution in E. coli . From the process, we discovered two mutants with the highest improvement (46 fold) in titer and the fastest k cat (44 fold) over the wild type 2PS, respectively, among those reported in the literature. A careful examination of the correlation between intracellular substrate concentration, Michaelis‐Menten parameters and TAL titer for these two mutants reveals that a fast reaction rate under limiting intracellular substrate concentrations is important for in‐cell biocatalysis. Such properties can be tuned by protein engineering and synthetic biology to adopt these engineered proteins for the maximum activities in different intracellular environments.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

The Importance of Nonequilibrium to Equilibrium Transition Pathways for the Efficiency and Stability of Organic Solar Cells

Controlled morphology of solution-processed thin films have realized impressive achievements for non-fullerene acceptor (NFA)–based organic solar cells (OSCs). Given the large set of donor–acceptor pairs, employing various processing conditions to realize optimal morphology for high efficiency and stable OSCs is a strenuous task. Therefore, comprehensive correlations between processing conditions and morphology evolution pathways have to be developed for efficient performance and stability of devices. Within the framework of the blend system, crystallization transitions of NFA molecules are tracked utilizing the first heating scan of differential scanning calorimeter (DSC) measurement correlating with respective morphology evolution of blend films. Real-time dynamics measurements and morphology characterizations are combined to provide optimal morphology transition pathways as NFA molecules are shown to be released from the mixed phase to form balanced ordered packing with variant processing conditions. Polymer:NFA films are fabricated using blade coating incorporating solvent additive or thermal annealing as processing conditions as a correlation is formulated between performance and stability of solar cells with morphology transition pathways. In conclusion, this work demonstrates the significance of processing condition-controlled transition pathways for the realization of optimal morphology leading to superior OSC devices.

14 SOLAR ENERGY↗

DOE OSTI2026

Out With the Old: Empirical Trends in U.S. Land‐Based Wind Turbine Decommissioning and Repowering

A growing number of wind turbines (WTs) across the globe are now reaching or exceeding their expected service lifetime; WT decommissioning is on the rise. Accordingly, questions pertaining to WT end-of-life have risen in importance in policy and practice. Yet, research on the various factors relating to WT decommissioning is relatively sparse. Moreover, the key assumptions underpinning that prior research (e.g., the lifespan of WTs, characteristics of WTs being decommissioned, and whether the site is repowered with new WTs) have never been empirically tested across a large set of decommissioned WTs. Leveraging a uniquely comprehensive and spatially explicit dataset of decommissioned WTs in the United States, this research analyzes spatial, technological, and temporal trends in WT decommissioning and develops a novel predictive model for WT decommissioning. Our analysis pinpoints more than 12,400 WTs that have been fully decommissioned in the United States., the majority of which have been relatively old (> 30 years) and small (< 200 kW). While a WT's age alone is a good predictor of the likelihood of decommissioning, other factors such as the size of the WT and recent performance are also important and significant predictors. Most sites where decommissioning has occurred have seen subsequent repowering, with repowered plants featuring substantially fewer WTs (−86 on average) and higher rated plant capacity (+62 MW on average). Many existing WTs in the U.S. are approaching the end of their expected life with roughly 7500 being 20 or more years old. Findings can help policymakers and stakeholders begin preparing for this potential wave of future decommissioning and repowering.

Decommissioning / End-of-life↗

DOE OSTI2025

Robustness of Deep Learning Classification to Adversarial Input on GPUs: Asynchronous Parallel Accumulation Is a Source of Vulnerability

The ability of machine learning (ML) classification models to resist small, targeted input perturbations—known as adversarial attacks—is a key measure of their safety and reliability. We show that floating-point non associativity (FPNA) coupled with asynchronous parallel programming on GPUs is sufficient to result in misclassification, without any perturbation to the input. Additionally, we show that this misclassification is particularly significant for inputs close to the decision boundary and that standard adversarial robustness results may be overestimated up to 4.6 when not considering machine-level details. We first study a linear classifier, before focusing on standard Graph Neural Network (GNN) architectures and datasets used in robustness assessments. We develop a novel black-box attack using Bayesian optimization to discover external workloads that can change the instruction scheduling which bias the output of reductions on GPUs and reliably lead to misclassification. Motivated by these results, we present a new learnable permutation (LP) gradient-based approach to learning floating-point operation orderings that lead to misclassifications. The LP approach provides a worst-case estimate in a computationally efficient manner, avoiding the need to run identical experiments tens of thousands of times over a potentially large set of possible GPU states or architectures. Finally, using instrumentation-based testing, we investigate parallel reduction ordering across different GPU architectures under external background workloads, when utilizing multi-GPU virtualization, and when applying power capping. Our results demonstrate that parallel reduction ordering varies significantly across architectures under the first two conditions, substantially increasing the search space required to fully test the effects of this parallel scheduler-based vulnerability. These results and the methods developed here can help to include machine-level considerations into adversarial robustness assessments, which can make a difference in safety and mission critical applications.

Shanmugavelu, Sanjif [Maxeler Technologies, a Groq↗

DOE OSTI2020

Lifting of level-1 states in the D1D5 CFT

The D1D5 CFT has a large set of states that are supersymmetric at the ‘free’ orbifold point in moduli space. When we perturb away from this point, some of these states join into long multiplets and lift in energy, while others remain supersymmetric. The count of unlifted states can be bounded below by an index, but the index does not yield the pattern of lifting; i.e., which states join into a long multiplet and how much this multiplet lifts. In this paper we consider the simple case of the D1D5 CFT where the orbifold CFT is a sigma model with targets space (T 4 ) 2 /S 2 and consider states at energy level 1. There are 2688 states at this level. The lifted states form a triplet of long multiplets, and we compute their lift at second order in perturbation theory. Half the members of the long multiplet are in the untwisted sector and half are in the twisted sector. This and other similar studies should help in the understanding of fuzzball states that describe extremal holes, since CFT sectors with low twist describe shallow throats in the dual gravity solution while sectors with high twist describe deep throats.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2021

Bottomonium suppression and elliptic flow using Heavy Quarkonium Quantum Dynamics

We introduce a framework called Heavy Quarkonium Quantum Dynamics (HQQD) which can be used to compute the dynamical suppression of heavy quarkonia propagating in the quark-gluon plasma using real-time in-medium quantum evolution. Using HQQD we compute large sets of real-time solutions to the Schrödinger equation using a realistic in-medium complex-valued potential. We sample 2 million quarkonia wave packet trajectories and evolve them through the QGP using HQQD to obtain their survival probabilities. The computation is performed using three different HQQD model parameter sets in order to estimate our systematic uncertainty. After taking into account final state feed down we compare our results to existing experimental data for the suppression and elliptic flow of bottomonium states and find that HQQD predictions are good agreement with available data for R AA as a function of N part and p T collected at \( \sqrt{s_{\mathrm{NN}}} \) = 5 . 02 TeV. In the case of v 2 for the various states, we find that the path-length dependence of ϒ(1 s ) suppression results in quite small v 2 for ϒ(1 s ). Our prediction for the integrated elliptic flow for ϒ(1 s ) in the 10 – 90% centrality class, which now includes an estimate of the systematic error, is v 2 [ϒ(1 s )] = 0 . 003 ± 0 . 0007 ± \( {}_{0.0013}^{0.0006} \) . We also find that, due to their increased suppression, excited bottomonium states have a larger elliptic flow. Based on this observation we make predictions for v 2 [ϒ(2 s )] and v 2 [ϒ(3 s )] as a function of centrality and transverse momentum.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗