Search NASA⌕ Search

SEARCH · Search NASA

Results for “ensemble optimization”

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 541 records · Page 30

Active operator learning with predictive uncertainty quantification for partial differential equations

With the increased prevalence of neural operators being used to provide rapid solutions to partial differential equations (PDEs), understanding the accuracy of model predictions and the associated error levels is necessary for deploying reliable surrogate models in scientific applications. Existing uncertainty quantification (UQ) frameworks employ ensembles or Bayesian methods, which can incur substantial computational costs during both training and inference. Here, we propose a lightweight predictive UQ method tailored for Deep operator networks (DeepONets) that also generalizes to other operator networks. Numerical experiments on linear and nonlinear PDEs demonstrate that the framework’s uncertainty estimates are unbiased and provide accurate out-of-distribution uncertainty predictions with a sufficiently large training dataset. Our framework provides fast inference and uncertainty estimates that can efficiently drive outer-loop analyses that would be prohibitively expensive with conventional solvers. We demonstrate how predictive uncertainties can be used in the context of Bayesian optimization and active learning problems to yield improvements in accuracy and data-efficiency for outer-loop optimization procedures. In the active learning setup, we extend the framework to Fourier Neural Operators (FNO) and describe a generalized method for other operator networks. To enable real-time deployment, we introduce an inference strategy based on precomputed trunk outputs and a sparse placement matrix, reducing evaluation time by more than a factor of five. Our method provides a practical route to uncertainty-aware operator learning in time-sensitive settings.

97 MATHEMATICS AND COMPUTING↗

Mapping the global distribution of C 4 vegetation using observations and optimality theory

Plants with the C 4 photosynthesis pathway typically respond to climate change differently from more common C 3 -type plants, due to their distinct anatomical and biochemical characteristics. These different responses are expected to drive changes in global C 4 and C 3 vegetation distributions. However, current C 4 vegetation distribution models may not predict this response as they do not capture multiple interacting factors and often lack observational constraints. Here, we used global observations of plant photosynthetic pathways, satellite remote sensing, and photosynthetic optimality theory to produce an observation-constrained global map of C 4 vegetation. We find that global C 4 vegetation coverage decreased from 17.7% to 17.1% of the land surface during 2001 to 2019. This was the net result of a reduction in C 4 natural grass cover due to elevated CO 2 favoring C 3 -type photosynthesis, and an increase in C 4 crop cover, mainly from corn (maize) expansion. Using an emergent constraint approach, we estimated that C 4 vegetation contributed 19.5% of global photosynthetic carbon assimilation, a value within the range of previous estimates (18–23%) but higher than the ensemble mean of dynamic global vegetation models (14 ± 13%; mean ± one standard deviation). Our study sheds insight on the critical and underappreciated role of C 4 plants in the contemporary global carbon cycle.

54 ENVIRONMENTAL SCIENCES↗

Basal Plane Doping to Activate Colloidal MoS 2 Nanosheets for Catalytic Hydrodeoxygenation of para- Cresol

The valorization of biomass into biofuels is a critical process for producing renewable fuels. Hydrodeoxygenation (HDO), particularly over doped molybdenum disulfide (MoS 2 ), a transition metal dichalcogenide (TMD) material, is a common representative catalytic reaction system for converting biomass-derived materials into useful hydrocarbons. However, the location and role of dopants, such as Co, in HDO is not fully understood. The effects of dopant location and oxidation state are often precluded by inhomogeneity in the ensemble properties of nanosheet size and dopant dispersion, as well as difficulty in observing the behavior of atomic site behavior directly. Using a colloidal approach to synthesize cobalt-doped MoS 2 nanosheets with controlled dopant concentration, combined with X-ray absorption spectroscopy (XAS) and density functional theory (DFT) calculations, we determine that basal plane doped Co (25% Co:Mo mole ratio) shows peak catalytic activity in HDO of para-cresol, a model biomass-derived compound, and that basal Co sites are demonstrably more active than edge sites. By observing these doping effects in MoS 2 catalysts for HDO, we can further optimize not only the production of carbon-neutral fuels but also direct the tailoring of doped TMD catalysts toward their intended applications.

36 MATERIALS SCIENCE↗

Dithering Strategies and Point-Source Photometry

The accuracy in the photometry of a point source depends on the point-spread function (PSF), detector pixelization, and observing strategy. The PSF and pixel response describe the spatial blurring of the source, the pixel scale describes the spatial sampling of a single exposure, and the observing strategy determines the set of dithered exposures with pointing offsets from which the source flux is inferred. In a wide-field imaging survey, sources of interest are randomly distributed within the field of view and hence are centered randomly within a pixel. A given hardware configuration and observing strategy therefore have a distribution of photometric uncertainty for sources of fixed flux that fall in the field. In this article we explore the ensemble behavior of photometric and position accuracies for different PSFs, pixel scales, and dithering patterns. We find that the average uncertainty in the flux determination depends slightly on dither strategy, whereas the position determination can be strongly dependent on the dithering. For cases with pixels much larger than the PSF, the uncertainty distributions can be non-Gaussian, with rms values that are particularly sensitive to the dither strategy. We also find that for these configurations with large pixels, pointings dithered by a fractional pixel amount do not always give minimal average uncertainties; this is in contrast to image reconstruction for which fractional dithers are optimal. When fractional pixel dithering is favored, a pointing accuracy of better than {approx}0.15 {approx}0.15 pixel width is required to maintain half the advantage over random dithers.

79 ASTRONOMY AND ASTROPHYSICS↗

Evaluating the Limits of QAOA Parameter Transfer at High-Rounds on Sparse Ising Models With Geometrically Local Cubic Terms

The emergent practical applicability of the Quantum Approximate Optimization Algorithm (QAOA) for approximate combinatorial optimization is a subject of considerable interest. One of the primary limitations of QAOA is the task of finding a set of good parameters, which is usually done using a variational optimization loop. Parameter transfer, or parameter concentration, is a phenomenon where QAOA angles trained on problem instances that are self-similar tend to perform well for other problem instances from that similar class. This suggests a potentially highly efficient and scalable non-variational learning method for QAOA angle finding. In this work, we systematically study QAOA parameter transferability from small problem sizes (16 and 27 decision variables) onto large problem instances (up to 156 qubits) for heavy-hex graph Ising models with geometrically local higher order terms using the Julia based QAOA simulation tool \texttt{JuliQAOA} to perform classical angle finding for up to $49$ QAOA layers ($p$). Parameter transfer of the fixed angles is validated using a combination of full statevector, Projected Entangled Pair States (PEPS), Matrix Product State (MPS), and LOWESA numerical simulations. We find that the QAOA parameter transfer from single instances applied to other (unseen) problem instances does not in general provide monotonically improving performance as a function of $p$ - there are many cases where the performance temporarily decreases as a function of $p$ - but despite this the transferred angles have a general trend of improved expectation value as the QAOA depth increases, in many cases converging close to the true ground-state energy of the $100+$ qubit instances. We also sample the hardware-compatible Ising models using the ensemble of transfer-learned QAOA parameters on several superconducting qubit IBM Quantum processors with 127, 133, and 156 qubits. We find continuous solution quality improvement of the hardware-compatible QAOA circuits run on the IBM NISQ processors up to $p=5$ on \texttt{ibm\_fez}, up to $p=9$ on \texttt{ibm\_torino}, and up to $p=10$ on \texttt{ibm\_pittsburgh}.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

How spin relaxes and dephases in bulk halide perovskites

Abstract Spintronics in halide perovskites has drawn significant attention in recent years, due to their highly tunable spin-orbit fields and intriguing interplay with lattice symmetry. Here, we perform first-principles calculations to determine the spin relaxation time ( T 1 ) and ensemble spin dephasing time ( $${T}_{2}^{*}$$ T 2 * ) in a prototype halide perovskite, CsPbBr 3 . To accurately capture spin dephasing in external magnetic fields we determine the Landé g -factor from first principles and take it into account in our calculations. These allow us to predict intrinsic spin lifetimes as an upper bound for experiments, identify the dominant spin relaxation pathways, and evaluate the dependence on temperature, external fields, carrier density, and impurities. We find that the Fröhlich interaction that dominates carrier relaxation contributes negligibly to spin relaxation, consistent with the spin-conserving nature of this interaction. Our theoretical approach may lead to new strategies to optimize spin and carrier transport properties.

74 ATOMIC AND MOLECULAR PHYSICS↗

Structural mechanisms for binding and activation of a contact-quenched fluorophore by RhoBAST

The fluorescent light-up aptamer RhoBAST, which binds and activates the fluorophore–quencher conjugate tetramethylrhodamine-dinitroaniline with high affinity, super high brightness, remarkable photostability, and fast exchange kinetics, exhibits excellent performance in super-resolution RNA imaging. Here we determine the co-crystal structure of RhoBAST in complex with tetramethylrhodamine-dinitroaniline to elucidate the molecular basis for ligand binding and fluorescence activation. The structure exhibits an asymmetric “A”-like architecture for RhoBAST with a semi-open binding pocket harboring the xanthene of tetramethylrhodamine at the tip, while the dinitroaniline quencher stacks over the phenyl of tetramethylrhodamine instead of being fully released. Molecular dynamics simulations show highly heterogeneous conformational ensembles with the contact-but-unstacked fluorophore–quencher conformation for both free and bound tetramethylrhodamine-dinitroaniline being predominant. The simulations also show that, upon RNA binding, the fraction of xanthene-dinitroaniline stacked conformation significantly decreases in free tetramethylrhodamine-dinitroaniline. This highlights the importance of releasing dinitroaniline from xanthene tetramethylrhodamine to unquench the RhoBAST–tetramethylrhodamine-dinitroaniline complex. Using SAXS and ITC, we characterized the magnesium dependency of the folding and binding mode of RhoBAST in solution and indicated its strong structural robustness. The structures and binding modes of relevant fluorescent light-up aptamers are compared, providing mechanistic insights for rational design and optimization of this important fluorescent light-up aptamer-ligand system.

59 BASIC BIOLOGICAL SCIENCES↗

AL4GAP: Active learning workflow for generating DFT-SCAN accurate machine-learning potentials for combinatorial molten salt mixtures

Machine learning interatomic potentials have emerged as a powerful tool for bypassing the spatiotemporal limitations of ab initio simulations, but major challenges remain in their efficient parameterization. We present AL4GAP, an ensemble active learning software workflow for generating multicomposition Gaussian approximation potentials (GAP) for arbitrary molten salt mixtures. The workflow capabilities include: (1) setting up user-defined combinatorial chemical spaces of charge neutral mixtures of arbitrary molten mixtures spanning 11 cations (Li, Na, K, Rb, Cs, Mg, Ca, Sr, Ba and two heavy species, Nd, and Th) and 4 anions (F, Cl, Br, and I), (2) configurational sampling using low-cost empirical parameterizations, (3) active learning for down-selecting configurational samples for single point density functional theory calculations at the level of Strongly Constrained and Appropriately Normed (SCAN) exchange-correlation functional, and (4) Bayesian optimization for hyperparameter tuning of two-body and many-body GAP models. Here, we apply the AL4GAP workflow to showcase high throughput generation of five independent GAP models for multicomposition binary-mixture melts, each of increasing complexity with respect to charge valency and electronic structure, namely: LiCl–KCl, NaCl–CaCl 2 , KCl–NdCl 3 , CaCl 2 –NdCl 3 , and KCl–ThCl 4 . Our results indicate that GAP models can accurately predict structure for diverse molten salt mixture with density functional theory (DFT)-SCAN accuracy, capturing the intermediate range ordering characteristic of the multivalent cationic melts.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

All-optical photoluminescence response of nitrogen-vacancy ensembles in diamond at low magnetic fields

All-optical (AO) microwave-free magnetometry using nitrogen-vacancy (N-𝑉) centers in diamond is attractive due to its broad applicability and reduced experimental complexity. In this work, we investigate room-temperature AO photoluminescence (PL) at low magnetic fields ( <2 mT) using diamonds with N-𝑉 ensembles at parts-per-million (ppm) concentrations. The measured AO-PL contrast features as a function of the applied magnetic field magnitude and direction are correlated with near-degenerate N-𝑉 electronic spin and hyperfine transitions from different N-𝑉 orientations within the diamond host. Reasonable agreement is found between low-field AO-PL measurements and model-based simulations of the effects of resonant dipolar interactions between N-𝑉 centers. The maximum observed AO-PL contrast depends on both the N-𝑉 concentration and the laser-illumination intensity at 532 nm. These results imply different optimal conditions for low-field AO N-𝑉 sensing compared to conventional optically detected magnetic resonance (ODMR) techniques, suggesting new research and application opportunities using AO measurements with lower system complexity, size, weight, and power.

Dipolar interaction↗

Out of Distribution Detection with Neural Network Anchoring

This is code to reproduce and build on OOD detection from the paper "Out of Distribution Detection with Neural Network Anchoring". Our goal here is to exploit heteroscedastic temperature scaling as a calibration strategy for out of distribution (OOD) detection. Heteroscedasticity here refers to the fact that the optimal temperature parameter for each sample can be different, as opposed to conventional approaches that use the same value for the entire distribution. To enable this, we propose a new training strategy called anchoring that can estimate appropriate temperature values for each sample, leading to state-of-the-art OOD detection performance across several benchmarks. Using NTK theory, we show that this temperature function estimate is closely linked to the epistemic uncertainty of the classifier, which explains its behavior. In contrast to some of the best-performing OOD detection approaches, our method does not require exposure to additional outlier datasets, custom calibration objectives, or model ensembling. Through empirical studies with different OOD detection settings - far OOD, near OOD, and semantically coherent OOD - we establish a highly effective OOD detection approach.

Thiagarajan, Jayaraman↗

ZettaBricks: A Language Compiler and Runtime System for Anyscale Computing

This grant supported the ZettaBricks and OpenTuner projects. ZettaBricks is a new implicitly parallel language and compiler where defining multiple implementations of multiple algorithms to solve a problem is the natural way of programming. ZettaBricks makes algorithmic choice a first class construct of the language. Choices are provided in a way that also allows our compiler to tune at a finer granularity. The ZettaBricks compiler autotunes programs by making both fine-grained as well as algorithmic choices. Choices also include different automatic parallelization techniques, data distributions, algorithmic parameters, transformations, and blocking. Additionally, ZettaBricks introduces novel techniques to autotune algorithms for different convergence criteria. When choosing between various direct and iterative methods, the ZettaBricks compiler is able to tune a program in such a way that delivers near-optimal efficiency for any desired level of accuracy. The compiler has the flexibility of utilizing different convergence criteria for the various components within a single algorithm, providing the user with accuracy choice alongside algorithmic choice. OpenTuner is a generalization of the experience gained in building an autotuner for ZettaBricks. OpenTuner is a new open source framework for building domain-specific multi-objective program autotuners. OpenTuner supports fully-customizable configuration representations, an extensible technique representation to allow for domain-specific techniques, and an easy to use interface for communicating with the program to be autotuned. A key capability inside OpenTuner is the use of ensembles of disparate search techniques simultaneously; techniques that perform well will dynamically be allocated a larger proportion of tests.

97 MATHEMATICS AND COMPUTING↗

Edge‐Driven Fringe‐Field Effects, Reduced Screening, and Bandgap Widening in Graphene Nanoribbons Enable Single‑Molecule Sensitivity

Graphene nanoribbons (GNRs) offer promising platforms for single‐molecule sensing due to their quasi‐1D channels and discrete electronic states, providing superior sensitivity toward molecular perturbations. While prior studies emphasize smoother edges as essential for optimal performance, the potential benefits of controlled edge roughness remain largely unexplored. Additionally, most investigations focus on GNR arrays, leaving critical edge‐ and width‐dependent factors, including fringe fields, bandgap widening, interactions between adsorbing molecules and GNR atoms, density of states (DOS) suppression, and electrostatic screening lengths, and their collective impact on sensitivity, poorly understood. Here, in this work, we fabricated field‐effect transistors using individual GNRs (widths: 200–20 nm) and characterized their response to molecular adsorption with perfluorooctanoic acid as the model analyte. Narrower ribbons displayed significantly enhanced sensitivity, yielding a coverage‐normalized response of 116 ± 10 mV per molecule in 20 nm‐wide GNRs (from calibrated ensemble Dirac‐point shifts). Experimental and theoretical analyses reveal that this heightened sensitivity arises from stronger fringe fields, width‐dependent quantum confinement effects, reduced DOS, and increased edge roughness that facilitates molecular anchoring, enhanced orbital overlap, and improved charge transfer efficiency. Our findings challenge the conventional assumption that smoother edges inherently enhance sensor performance, demonstrating that controlled edge disorder substantially boosts molecular sensitivity in GNR sensors.

36 MATERIALS SCIENCE↗

Discovery of Multiple Light-Harvesting States of the Photosynthetic Protein PE545

Cryptophytes are photosynthetic microalga that flourish in a remarkable diversity of natural environments by using pigment-containing proteins with absorption maxima tuned to each ecological niche. While this diversity in the absorption has been well established, the subsequent photophysics is highly sensitive to the local protein environment and so may exhibit similar variation. Thermal fluctuations of the protein conformation are expected to introduce photophysical heterogeneity of the pigments that may have evolved important functional properties in a manner similar to that of the absorption. However, such heterogeneity is averaged out in ensemble measurements and, therefore, has not yet been probed. Here, we report single-molecule measurements of phycoerythrin 545 (PE545), the prototypical cryptophyte antenna protein, in its native dimeric form. A conformational ensemble was resolved consisting of distinct photophysical states with different light-harvesting properties. Proteins that did not quench, partially quenched, or fully quenched absorbed light were observed. Light intensity increased the quenched-state population of the dimer, potentially as a mechanism to deal with the extreme light intensities found in aqueous environments. Cross-linking, which mimics local interactions, introduces this light-dependent functionality while also suppressing other conformational dynamics. The cellular organization can, therefore, actively modulate the protein conformation and dynamics, selecting for distinct levels of light harvesting. Furthermore, the complex conformational equilibrium provides an additional mechanism for cryptophytes and likely other photosynthetic organisms to optimize solar energy capture and conversion.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Hybrid quantum-classical algorithms for approximate graph coloring

We show how to apply the recursive quantum approximate optimization algorithm (RQAOA) to MAX- k -CUT, the problem of finding an approximate k -vertex coloring of a graph. We compare this proposal to the best known classical and hybrid classical-quantum algorithms. First, we show that the standard (non-recursive) QAOA fails to solve this optimization problem for most regular bipartite graphs at any constant level p : the approximation ratio achieved by QAOA is hardly better than assigning colors to vertices at random. Second, we construct an efficient classical simulation algorithm which simulates level- 1 QAOA and level- 1 RQAOA for arbitrary graphs. In particular, these hybrid algorithms give rise to efficient classical algorithms, and no benefit arising from the use of quantum mechanics is to be expected. Nevertheless, they provide a suitable testbed for assessing the potential benefit of hybrid algorithm: We use the simulation algorithm to perform large-scale simulation of level- 1 QAOA and RQAOA with up to 300 qutrits applied to ensembles of randomly generated 3 -colorable constant-degree graphs. We find that level- 1 RQAOA is surprisingly competitive: for the ensembles considered, its approximation ratios are often higher than those achieved by the best known generic classical algorithm based on rounding an SDP relaxation. This suggests the intriguing possibility that higher-level RQAOA may be a potentially useful algorithm for NISQ devices.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Effects of climate change on streamflow extremes and implications for reservoir inflow in the United States

The magnitude and frequency of hydrometeorological extremes are expected to increase in the conterminous United States (CONUS) over the rest of this century, and their increase will significantly impact water resource management. While previous efforts focused on the effects of reservoirs on downstream discharge, the effects of climate change on reservoir inflows in upstream areas are not well understood. We evaluated the large-scale climate change effects on extreme hydrological events and their implications for reservoir inflows in 178 headwater basins across CONUS using the Variable Infiltration Capacity (VIC) hydrologic model. The VIC model was forced with a 10-member ensemble of global circulation models under the Representative Concentration Pathway 8.5 that were dynamically downscaled using a regional climate model (RegCM4) and bias-corrected to 1/24° grid cell resolution. The results projected an increase in the likelihood of flood risk by 44% for a majority of subbasins upstream of flood control reservoirs in the central United States and increased drought risk by 11% for subbasins upstream of hydropower reservoirs across the western United States. Increased risk of both floods and droughts can potentially make reservoirs across CONUS more vulnerable to future climate conditions. In conclusion, this study estimates reservoir inflow changes over the next several decades, which can be used to optimize water supply management downstream.

54 ENVIRONMENTAL SCIENCES↗

Finite-Temperature Structures of Supported Subnanometer Catalysts Inferred via Statistical Learning and Genetic Algorithm-Based Optimization

Single-atom catalysts (SACs) minimize noble metal utilization and can alter the activity and selectivity of supported metal nanoparticles. However, the morphology of active centers, including single atoms and subnanometer clusters of a few atoms, remains elusive due to experimental challenges. The computational cost to describe numerous cluster shapes and sizes makes direct first-principles calculations impractical. We present a computational framework to enable structure determination for single-atom and subnanometer cluster catalysts. As a case study, we obtained the low energy structures of Pd n (n = 1-21) clusters supported on CeO 2 (111), which are critical components of automobile three-way catalysts. Trained on density functional theory data, a three-dimensional cluster expansion is established using statistical learning to describe the Hamiltonian and predict energies of supported Pdn clusters of any structure. Low energy stable and metastable structures are identified using a Metropolis Monte Carlo-based genetic algorithm in the canonical ensemble at 300 K. We observe that supported single atoms sinter to form bilayer clusters and large cluster isomers share similarities in both shape and energy, and elucidate the significance of the support and microstructure on cluster stability. We discovered a simple surrogate structure-energy model, where the energy per atom scales with the square root of the average first coordination number, which can be used to estimate energies and compare the stability of clusters. Our framework, applicable to any metal/support system, fills an important methodological gap to predict the stability of supported metal catalysts in the subnanometer regime.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Measuring, Quantifying, and Predicting the Cost-Accuracy Tradeoff

Exponentially increasing data volumes, coupled with new modes of analysis have created significant new opportunities for data scientists. However, the stochastic nature of many data science techniques results in tradeoffs between costs and accuracy. For example, machine learning algorithms can be trained iteratively and indefinitely with diminishing returns in terms of accuracy. In this paper we explore the cost-accuracy tradeoff through three representative examples: we vary the number of models in an ensemble, the number of epochs used to train a machine learning model, and the amount of data used to train a machine learning model. We highlight the feasibility and benefits of being able to measure, quantify, and predict cost-accuracy tradeoffs by demonstrating the presence and usability of these tradeoffs in two different case studies.

big data analytics↗

Deep Learning Image Segmentation for Atmospheric Rivers

Abstract The identification of atmospheric rivers (ARs) is crucial for weather and climate predictions as they are often associated with severe storm systems and extreme precipitation, which can cause large impacts on society. This study presents a deep learning model, termed ARDetect, for image segmentation of ARs using ERA5 data from 1960 to 2020 with labels obtained from the TempestExtremes tracking algorithm. ARDetect is a convolutional neural network (CNN)-based U-Net model, with its structure having been optimized using automatic hyperparameter tuning. Inputs to ARDetect were selected to be the integrated water vapor transport (IVT) and total column water (TCW) fields, as well as the AR mask from TempestExtremes from the previous time step to the one being considered. ARDetect achieved a mean intersection-over-union (mIoU) rate of 89.04% for ARs, indicating its high accuracy in identifying these weather patterns and a superior performance than most deep learning–based models for AR detection. In addition, ARDetect can be executed faster than the TempestExtremes method (seconds vs minutes) for the same period. This provides a significant benefit for online AR detection, especially for high-resolution global models. An ensemble of 10 models, each trained on the same dataset but having different starting weights, was used to further improve on the performance produced by ARDetect, thus demonstrating the importance of model diversity in improving performance. ARDetect provides an effective and fast deep learning–based model for researchers and weather forecasters to better detect and understand ARs, which have significant impacts on weather-related events such as floods and droughts.

Galea, Daniel↗