Search NASA⌕ Search

SEARCH · Search NASA

Results for “distribution shift”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Assessing Membership Inference Attacks under Distribution Shifts

Membership inference attacks (MIAs) exploit machine learning models to infer whether a data point was in the training set, posing significant privacy risks even with limited black-box access. These attacks rely on the attacker approximating the target model’s training distribution, yet the impact of distribution shifts between target and shadow models on MIA success remains underexplored. We systematically evaluate five types of distribution shifts —-cutout, jitter, Gaussian noise, label shift, and attribute shift —- at varying intensities. Our results reveal that these shifts affect MIA effectiveness in nuanced ways, with some reducing attack success while others exacerbate vulnerabilities, and the same shift can have opposite effects depending on the type of MIA. This highlights the complex interplay between distributional differences and attack performance, offering critical insights for improving model defenses against MIAs.

Shi, Yichuan [Massachusetts Institute of Technolog↗
DOE OSTI2019

Robust, automated sleep scoring by a compact neural network with distributional shift correction

Studying the biology of sleep requires the accurate assessment of the state of experimental subjects, and manual analysis of relevant data is a major bottleneck. Recently, deep learning applied to electroencephalogram and electromyogram data has shown great promise as a sleep scoring method, approaching the limits of inter-rater reliability. As with any machine learning algorithm, the inputs to a sleep scoring classifier are typically standardized in order to remove distributional shift caused by variability in the signal collection process. However, in scientific data, experimental manipulations introduce variability that should not be removed. For example, in sleep scoring, the fraction of time spent in each arousal state can vary between control and experimental subjects. We introduce a standardization method, mixture z-scoring, that preserves this crucial form of distributional shift. Using both a simulated experiment and mouse in vivo data, we demonstrate that a common standardization method used by state-of-the-art sleep scoring algorithms introduces systematic bias, but that mixture z-scoring does not. We present a free, open-source user interface that uses a compact neural network and mixture z-scoring to allow for rapid sleep scoring with accuracy that compares well to contemporary methods. This work provides a set of computational tools for the robust automation of sleep scoring.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Data-Conforming Data-Driven Control: Avoiding Premature Generalizations Beyond Data

Data-driven and adaptive control approaches face the problem of introducing sudden distributional shifts beyond the distribution of data encountered during learning. Therefore, they are prone to invalidating the very assumptions used in their own construction. This is due to the linearity of the underlying system, inherently assumed and formulated in most data-driven control approaches, which may falsely generalize the behavior of the system beyond the behavior experienced in the data. This article seeks to mitigate these problems by enforcing consistency of the newly designed closed-loop systems with data and slowing down any distributional shifts in the joint state-input space. This is achieved through incorporating affine regularization terms and linear matrix inequality constraints to data-driven approaches, resulting in convex semi-definite programs that can be efficiently solved by standard software packages. We discuss the optimality conditions of these programs and then conclude this article with a numerical example that further highlights the problem of premature generalization beyond data and shows the effectiveness of our proposed approaches in enhancing the safety of data-driven control methods.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Domain-Adapted Convolutional Networks for Satellite Image Classification: A Large-Scale Interactive Learning Workflow

Satellite imagery often exhibits large spatial extent areas that encompass object classes with considerable variability. This often limits large-scale model generalization with machine learning algorithms. Notably, acquisition conditions, including dates, sensor position, lighting condition, and sensor types, often translate into class distribution shifts introducing complex nonlinear factors and hamper the potential impact of machine learning classifiers. Here, this article investigates the challenge of exploiting satellite images using convolutional neural networks (CNN) for settlement classification where the class distribution shifts are significant. We present a large-scale human settlement mapping workflow based-off multiple modules to adapt a pretrained CNN to address the negative impact of distribution shift on classification performance. To extend a locally trained classifier onto large spatial extents areas we introduce several submodules: First, a human-in-the-loop element for relabeling of misclassified target domain samples to generate representative examples for model adaptation; second, an efficient hashing module to minimize redundancy and noisy samples from the mass-selected examples; and third, a novel relevance ranking module to minimize the dominance of source example on the target domain. The workflow presents a novel and practical approach to achieve large-scale domain adaptation with binary classifiers that are based-off CNN features. Experimental evaluations are conducted on areas of interest that encompass various image characteristics, including multisensors, multitemporal, and multiangular conditions. Domain adaptation is assessed on source–target pairs through the transfer loss and transfer ratio metrics to illustrate the utility of the workflow.

47 OTHER INSTRUMENTATION↗
DOE OSTI2016

How uncertain is the future of electric vehicle market: Results from Monte Carlo simulations using a nested logit model

Plug-in electric vehicles (PEVs) are widely regarded as an important component of the technology portfolio designed to accomplish policy goals in sustainability and energy security. However, the market acceptance of PEVs in the future remains largely uncertain from today's perspective. By integrating a consumer choice model based on nested multinomial logit and Monte Carlo simulation, this study analyzes the uncertainty of PEV market penetration using Monte Carlo simulation. Results suggest that the future market for PEVs is highly uncertain and there is a substantial risk of low penetration in the early and midterm market. Top factors contributing to market share variability are price sensitivities, energy cost, range limitation, and charging availability. The results also illustrate the potential effect of public policies in promoting PEVs through investment in battery technology and infrastructure deployment. Here, continued improvement of battery technologies and deployment of charging infrastructure alone do not necessarily reduce the spread of market share distributions, but may shift distributions toward right, i.e., increase the probability of having great market success.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2015

An efficient reconstruction algorithm for differential phase-contrast tomographic images from a limited number of views

The main focus of this paper is reconstruction of tomographic phase-contrast image from a set of projections. We propose an efficient reconstruction algorithm for differential phase-contrast computed tomography that can considerably reduce the number of projections required for reconstruction. The key result underlying this research is a projection theorem that states that the second derivative of the projection set is linearly related to the Laplacian of the tomographic image. The proposed algorithm first reconstructs the Laplacian image of the phase-shift distribution from the second-derivative of the projections using total variation regularization. The second step is to obtain the phase-shift distribution by solving a Poisson equation whose source is the Laplacian image previously reconstructed under the Dirichlet condition. We demonstrate the efficacy of this algorithm using both synthetically generated simulation data and projection data acquired experimentally at a synchrotron. The experimental phase data were acquired from a human coronary artery specimen using dark-field-imaging optics pioneered by our group. Our results demonstrate that the proposed algorithm can reduce the number of projections to approximately 33% as compared with the conventional filtered backprojection method, without any detrimental effect on the image quality.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

A biology-informed similarity metric for simulated patches of human cell membrane

Complex scientific inquiries rely increasingly upon large and autonomous multiscale simulation campaigns, which fundamentally require similarity metrics to quantify ‘sufficient’ changes among data and/or configurations. However, subject matter experts are often unable to articulate similarity precisely or in terms of well-formulated definitions, especially when new hypotheses are to be explored, making it challenging to design a meaningful metric. Furthermore, the key to practical usefulness of such metrics to enable autonomous simulations lies in in situ inference, which requires generalization to possibly substantial distributional shifts in unseen, future data. Here, we address these challenges in a cancer biology application and develop a meaningful similarity metric for ‘patches’—regions of simulated human cell membrane that express interactions between certain proteins of interest and relevant lipids. In the absence of well-defined conditions for similarity, we leverage several biology-informed notions about data and the underlying simulations to impose inductive biases on our metric learning framework, resulting in a suitable similarity metric that also generalizes well to significant distributional shifts encountered during the deployment. We combine these intuitions to organize the learned embedding space in a multiscale manner, which makes the metric robust to incomplete and even contradictory intuitions. Our approach delivers a metric that not only performs well on the conditions used for its development and other relevant criteria, but also learns key spatiotemporal relationships without ever being exposed to any such information during training.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Multiscale Flow for robust and optimal cosmological analysis

We propose Multiscale Flow, a generative Normalizing Flow that creates samples and models the field-level likelihood of two-dimensional cosmological data such as weak lensing. Multiscale Flow uses hierarchical decomposition of cosmological fields via a wavelet basis and then models different wavelet components separately as Normalizing Flows. The log-likelihood of the original cosmological field can be recovered by summing over the log-likelihood of each wavelet term. This decomposition allows us to separate the information from different scales and identify distribution shifts in the data such as unknown scale-dependent systematics. The resulting likelihood analysis can not only identify these types of systematics, but can also be made optimal, in the sense that the Multiscale Flow can learn the full likelihood at the field without any dimensionality reduction. We apply Multiscale Flow to weak lensing mock datasets for cosmological inference and show that it significantly outperforms traditional summary statistics such as power spectrum and peak counts, as well as machine learning–based summary statistics such as scattering transform and convolutional neural networks. We further show that Multiscale Flow is able to identify distribution shifts not in the training data such as baryonic effects. Finally, we demonstrate that Multiscale Flow can be used to generate realistic samples of weak lensing data.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2013

First observation of the Λ ( 1405 ) line shape in electroproduction

We report the first observation of the line shape of the Λ ( 1405 ) from electroproduction, and show that it is not a simple Breit-Wigner resonance. Electroproduction of K + Λ ( 1405 ) off the proton was studied by using data from CLAS at Jefferson Lab in the range 1.0 < Q 2 < 3.0 (GeV/ c ) 2 . The analysis utilized the decay channels Σ + π - of the Λ ( 1405 ) and p π 0 of the Σ + . Neither the standard Particle Data Group resonance parameters, nor free parameters fitting to a single Breit-Wigner resonance represent the line shape. In our fits, the line shape corresponds approximately to predictions of a two-pole meson-baryon picture of the Λ ( 1405 ) , with a lower mass pole near 1368 MeV/ c 2 and a higher mass pole near 1423 MeV/ c 2 . Furthermore, with increasing photon virtuality the mass distribution shifts toward the higher mass pole.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Methods for Continuously Resolving Spectral Shifts in Distributed Optical Fiber Sensors Irradiated to Extreme Neutron Fluence

Optical frequency domain reflectometry (OFDR) is a technique for interrogating distributed optical fiber sensors (DOFS) and involves correlating changes in the Rayleigh backscatter fingerprint for a fiber under test (FUT) with a reference measurement. Recently, under the WIRE-21 experiment sponsored by Nuclear Science User Facilities (NSUF) and performed at the High-Flux Isotope Reactor (HFIR) at Oak Ridge National Laboratory (ORNL), silica optical fibers were irradiated to a fast neutron fluence of 1x10^21 n/cm^2/s at temperatures from 200--400 C. As in the cases of high-temperature and high-strain conditions, such high levels of neutron bombardment result in a highly dynamic RBS that evades analysis with conventional methods. This work describes the further development and application of graphical signal processing techniques applied to OFDR-based distributed optical fiber sensors specifically deployed in in-pile applications. While the signal processing techniques developed in this work are applied to DOFS in nuclear environments, they provide a general framework for the analysis of OFDR measurements and a tangible method to yield higher quality data, without imposing additional hardware requirements.

Sweeney, Dan↗
DOE OSTI2015

Electric wind in a Differential Mobility Analyzer

Electric wind -- the movement of gas, induced by ions moving in an electric field -- can be a distorting factor in size distribution measurements using Differential Mobility Analyzers (DMAs). The aim of this study was to determine the conditions under which electric wind occurs in the locally-built VLDMA (Very Long Differential Mobility Analyzer) and TSI Long-DMA (3081) and to describe the associated distortion of the measured spectra. Electric wind proved to be promoted by the increase of electric field strength, aerosol layer thickness, particle number concentration and particle size. The measured size spectra revealed three types of distortion: widening of the size distribution, shift of the mode of the distribution to smaller diameters and smoothing out the peaks of the multiply charged particles. Electric wind may therefore be a source of severe distortion of the spectrum when measuring large particles at high concentrations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Short-term Electricity Price Forecasting with Constrained Regressors

The volatility of electricity price presents a challenge to market participants as their decision-making process are highly depend on the accuracy of price forecasts. However, there is growing empirical evidence of increasing price volatility and price spikes in electricity markets as a result of variable renewable energy generation, extreme weather events, and other factors. The distribution shift caused by spikes in electricity price data differentiates the forecasting tasks from other renewable energy sources. Moreover, the observations may be compromised by cyberattacks and thus not available in the testing phase. To this end, we propose a Similarity-Enhanced Electricity Decomposition Forecasting model (SEED-Forecaster) to address the missing response problem and spikes capturing in short-term electricity price forecasting. The effectiveness of the proposed framework is tested on real-world electricity price data from California Independent System Operator (CAISO). Numerical results of case studies show that the proposed SEED-Forecsater can enhance forecasting performance, particularly in capturing electricity spikes, even under conditions without regressors during testing stage.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2012

YPdSn and YPd{sub 2}Sn: Structure, {sup 89}Y solid state NMR and {sup 119}Sn Moessbauer spectroscopy

The stannides YPdSn and YPd{sub 2}Sn were synthesized by high-frequency melting of the elements in sealed tantalum tubes. Both structures were refined on the basis of single crystal X-ray diffractometer data: TiNiSi type, Pnma, a=715.4(1), b=458.8(1), c=789.1(1) pm, wR2=0.0461, 510 F{sup 2} values, 20 variables for YPdSn and MnCu{sub 2}Al type, Fm3 Macron m, a=671.44(8), wR2=0.0740, 55 F{sup 2} values, 5 parameters for YPd{sub 2}Sn. The yttrium atoms in the new stannide YPdSn are coordinated by two tilted Pd{sub 3}Sn{sub 3} hexagons (ordered AlB{sub 2} superstructure). In the Heusler phase YPd{sub 2}Sn each yttrium atom has octahedral tin coordination and additionally eight palladium neighbors. The cubic site symmetry of yttrium is reflected in the {sup 119}Sn Moessbauer spectrum which shows no quadrupole splitting. In contrast, YPdSn shows a single signal at {delta}=1.82(1) mm/s subjected to quadrupole splitting of {Delta}E{sub Q}=0.93(1) mm/s. Both compounds have been characterized by high-resolution {sup 89}Y solid state NMR spectroscopy, which indicates the presence of strong Knight shifts. The spectrum of YPd{sub 2}Sn is characterized by an unusually large linewidth, suggesting the presence of a Knight shift distribution reflecting local disordering effects. The range of {sup 89}Y Knight shifts of several binary and ternary intermetallic yttrium compounds is briefly discussed. - Graphical abstract: YPdSn and YPd{sub 2}Sn: Structure, {sup 89}Y solid state NMR and {sup 119}Sn Moessbauer spectroscopy. Highlights: Black-Right-Pointing-Pointer Synthesis and structure of ternary stannides YPdSn and YPd{sub 2}Sn. Black-Right-Pointing-Pointer {sup 119}Sn Moessbauer spectroscopic investigation of YPdSn and YPd{sub 2}Sn. Black-Right-Pointing-Pointer {sup 89}Y solid state NMR of intermetallics.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2017

Brazil-USA Collaborative Research: Modifications by Anthropogenic Pollution of the Natural Atmospheric Chemistry and Particle Microphysics of the Tropical Rain Forest During the GoAmazon Intensive Operating Periods (IOPs)

Manaus, a city of nearly two million people, represents an isolated urban area having a distinct urban pollution plume within the otherwise pristine Amazon Basin. The plume has high concentrations of oxides of nitrogen and sulfur, carbon monoxide, particle concentrations, and soot, among other pollutants. Critically, the distinct plume in the setting of the surrounding tropical rain forest serves as a natural laboratory to allow direct comparisons between periods of pollution influence to those of pristine conditions. The funded activity of this report is related to the Brazil-USA collaborative project during the two Intensive Operating Periods (wet season, 1 Feb - 31 Mar 2014; dry season, 15 Aug - 15 Oct 2014) of GoAmazon2014/5. The project addresses key science questions regarding the modification of the natural atmospheric chemistry and particle microphysics of the forest by present and future anthropogenic pollution. The first objective of the project was to understand and quantify the interactions of biogenic and anthropogenic emissions with respect to the production of secondary organic material. In clean conditions in the Amazon basin, secondary organic material dominates the diameter distribution of the submicron particles. How and why is the diameter distribution shifted by pollution? The second objective followed from the first in that, although the diameter distribution is dominated by secondary organic material, the actual source of new particle production remains uncertain (i.e., the number concentration). The second objective was to test the hypothesis that new particles under natural conditions are produced as a result of evaporation of primary particles emitted by fungal spores as well as to investigate any shifts in this mechanism under pollution conditions, e.g., in consequence to the high concentrations of SO 2 in the pollution plume. Combined, the number-diameter distribution is the key connection to upscaling to the effects of aerosol particles on clouds and climate. Understanding this upscaling under clean and pollution conditions, including differences, was the third objective.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Evaluating Entrainment–Mixing Characteristics through Direct Comparisons of Drop Size Distributions Using In Situ Observations from ACE-ENA

Abstract Constraining the impacts of entrainment and associated mixing (i.e., entrainment–mixing) on cloud properties continues to be difficult, partly due to observational uncertainties as well as a lacking consensus of which methodologies for diagnosing entrainment–mixing are most appropriate. This study introduces a novel method to evaluate the presence and degree of inhomogeneous and homogeneous mixing using ∼100 h of in situ observations from a research aircraft over the northeastern Atlantic. Specifically, drop size distributions are compared between regions containing negligible and significant entrainment for select flight legs, making a direct characterization of the degree of homogeneous and inhomogeneous mixing possible. A measure of drop concentration variance is used as a proxy variable to diagnose entrainment–mixing. Results correspond well with entrainment–mixing metrics, showing lower Damköhler numbers where drop size distributions shift toward smaller drop sizes (i.e., inhomogeneous mixing) and greater transition length scales where drop size distributions do not (i.e., homogeneous mixing). Inhomogeneous mixing occurs in most samples from Aerosol and Cloud Experiment in the Eastern North Atlantic (ACE-ENA) (regardless of homogeneous mixing frequencies increasing with increasing spatial resolution from ∼100 to ∼10 m) and is associated with decreased drop size relative dispersion and both greater aerosol and drop concentrations compared with homogeneous mixing. Precipitating clouds have a greater frequency of homogeneous mixing compared with nonprecipitating clouds. The proposed methodology is similarly applied to in situ observations of southeast Pacific stratocumulus, shallow convective clouds over central Oklahoma and low-level clouds over the Southern Ocean. All four locations are primarily dominated by inhomogeneous mixing with minimal variability among each region.

Clouds↗
DOE OSTI2016

Generating multiple orbital angular momentum vortex beams using a metasurface in radio frequency domain

In this paper, an electromagnetic metasurface is designed, fabricated, and experimentally demonstrated to generate multiple orbital angular momentum (OAM) vortex beams in radio frequency domain. Theoretical formula of compensated phase-shift distribution is deduced and used to design the metasurface to produce multiple vortex radio waves in different directions with different OAM modes. The prototype of a practical configuration of square-patch metasurface is designed, fabricated, and measured to validate the theoretical analysis at 5.8 GHz. The simulated and experimental results verify that multiple OAM vortex waves can be simultaneously generated by using a single electromagnetic metasurface. The proposed method paves an effective way to generate multiple OAM vortex waves in radio and microwave wireless communication applications.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗