SEARCH · Search NASA
Results for “Deep space”
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.
PRIMED for the Future: Purposing Raw Intake for Machine Learning-Enabled Detection (Final Report)
The COVID-19 pandemic demonstrated how a novel, elusive, and diffuse biological threat can engender uncertainty and misinformation, and it underscored the need for flexible analytical modalities agnostic to the identity of biological material. Yet even before the pandemic, recognition of the limitations of the current, list-based approach, which focuses on known pathogens and biotoxins, and of the importance of agent-agnostic biodetection was growing within the biosecurity community. In a 2018 report on “Biodefense in the Age of Synthetic Biology,” for example, the National Academy of Sciences stated that “an overreliance on the Select Agent List is a systematic weakness affecting many aspects of the United States’ current biodefense mitigation capability”. More recently, a group of biodefense researchers proposed the identification and adoption of “bioagent-agnostic signatures (BASs)” as a way of detecting and characterizing not only existing agents but also novel ones, an approach they believe will “enable a more flexible and resilient biodefense posture”. Indeed, the future of biodetection requires us to begin developing novel analytics that can identify anomalies and/or characteristics that indicate a potential threat, whether known or unknown, without looking for a specific signature that has been identified previously. To assess potential threats more rapidly, it is critical to develop agnostic artificial intelligence (AI)/machine learning (ML) systems that can be employed for real-time assessment of the nature and source of a perturbation. Such systems should be multi scale and multi-dimensional, integrating sensor data from a range of biological, chemical, and physical application spaces. Emerging deep learning (DL) models demonstrate exceptional promise for identification of discriminatory features within multi-dimensional datasets. DL models have the capacity to recognize and encode highly complex patterns in a wide range of input data modalities, including images, text, and biological/chemical/physical spectra. As such, they can execute a wide range of assessments and determinations that have traditionally required a human operator. The promise of advances in DL is apparent in the realm of human health and medicine. DL models have been validated for evaluating a variety of clinical threats to human health in a range of contexts, including infection and cancer, and they demonstrated improved performance in predicting stroke relative to human neurologists in some categories of data. Continuously evolving advances in AI/ML are expected to support more efficient evaluation of raw sequence, spectroscopy, and spectrometry data. For instance, recent advances and deployment of large language models (LLM) such as Generative Pre training Transformer (GPT) and Bidirectional Encoder Representations from Transformers (BERT) have already motivated application of these models for biological function prediction. As frameworks such as LLMs become larger and more complex in their representations, their capacity to serve as pre-trained models that can be fine-tuned for biological/biodetection purposes will similarly be amplified. While existing and emerging AI/ML have found broad applicability and use cases in the clinical sciences, development for environmental evaluation and biodetection has been limited. Functionalizing such capabilities for this purpose requires an understanding of the existing technical landscape and how the respective tools and algorithms are currently being employed. This landscape awareness then allows an assessment of the current practical capabilities of existing models and the anticipated requirements and development efforts that will be needed to adapt available algorithms for biodetection applications relevant to DHS. Leveraging expertise in biodetection, ML, and operational biodetection, the effort described in this report is comprised of a systematic landscape assessment (Subtask 2.1), comparative evaluation (Subtask 2.2), and formulation of a value proposition (Subtask 2.3) for the prospect of ML-enabled, agnostic biodetection from raw, or minimally-processed, datasets.
PRIMED for the Future: Purposing Raw Intake for Machine Learning-Enabled Detection
The COVID-19 pandemic demonstrated how a novel, elusive, and diffuse biological threat can engender uncertainty and misinformation, and it underscored the need for flexible analytical modalities agnostic to the identity of biological material. Yet even before the pandemic recognition of the limitations of the current, list-based approach, which focuses on known pathogens and biotoxins, and of the importance of agent-agnostic biodetection, was growing within the biosecurity community. In a 2018 report on “Biodefense in the Age of Synthetic Biology,” for example, the National Academy of Sciences stated that “an overreliance on the Select Agent List is a systematic weakness affecting many aspects of the United States’ current biodefense mitigation capability." More recently, a group of biodefense researchers proposed the identification and adoption of “bioagent-agnostic signatures (BASs)” as a way of detecting and characterizing not only existing agents but also novel ones, an approach they believe will “enable a more flexible and resilient biodefense posture." Indeed, the future of biodetection requires us to begin developing novel analytics that can identify anomalies and/or characteristics that indicate a potential threat, whether known or unknown, without looking for a specific signature that has been identified previously. To assess potential threats more rapidly, it is critical to develop agnostic artificial intelligence (AI)/machine learning (ML) systems that can be employed for real-time assessment of the nature and source of a perturbation. Such systems should be multiscale and multi-dimensional, integrating sensor data from a range of biological, chemical, and physical application spaces. Emerging deep learning (DL) models demonstrate exceptional promise for identification of discriminatory features within multi-dimensional datasets. DL models have the capacity to recognize and encode highly complex patterns in a wide range of input data modalities, including images, text, and biological/chemical/physical spectra. As such, they can execute a wide range of assessments and determinations that have traditionally required a human operator.
Eridanus II: A Fossil from Reionization with an Off-center Star Cluster
In this work, we present deep Hubble Space Telescope (HST) photometry of the ultra-faint dwarf galaxy Eridanus II (Eri II). Eri II, which has an absolute magnitude of M V = -7.1, is located at a distance of 339 kpc, just beyond the virial radius of the Milky Way. We determine the star formation history of Eri II and measure the structure of the galaxy and its star cluster. We find that a star formation history consisting of two bursts, constrained to match the spectroscopic metallicity distribution of the galaxy, accurately describes the Eri II stellar population. The best-fit model implies a rapid truncation of star formation at early times, with >80% of the stellar mass in place before z ~ 6. A small fraction of the stars could be as young as 8 Gyr, but this population is not statistically significant; Monte Carlo simulations recover a component younger than 9 Gyr only 15% of the time, where they represent an average of 7 ± 4% of the population. These results are consistent with theoretical expectations for quenching by reionization. The HST depth and angular resolution enable us to show that Eri II’s cluster is offset from the center of the galaxy by a projected distance of 23 ± 3 pc. This offset could be an indication of a small (~50–75 pc) dark matter core in Eri II. Moreover, we demonstrate that the cluster has a high ellipticity of and is aligned with the orientation of Eri II within 3° ± 6°, likely due to tides. The stellar population of the cluster is indistinguishable from that of Eri II itself.
Timing the r -process Enrichment of the Ultra-faint Dwarf Galaxy Reticulum II
The ultra-faint dwarf galaxy Reticulum II (Ret II) exhibits a unique chemical evolution history, with ${72}_{-12}^{+10}$% of its stars strongly enhanced in r-process elements. We present deep Hubble Space Telescope photometry of Ret II and analyze its star formation history. As in other ultra-faint dwarfs, the color–magnitude diagram is best fit by a model consisting of two bursts of star formation. If we assume that the bursts were instantaneous, then the older burst occurred around the epoch of reionization, forming ~80% of the stars in the galaxy, while the remainder of the stars formed ~3 Gyr later. When the bursts are allowed to have nonzero durations, we obtain slightly better fits. The best-fitting model in this case consists of two bursts beginning before reionization, with approximately half the stars formed in a short (100 Myr) burst and the other half in a more extended period lasting 2.6 Gyr. Considering the full set of viable star formation history models, we find that 28% of the stars formed within 500 ± 200 Myr of the onset of star formation. The combination of the star formation history and the prevalence of r-process-enhanced stars demonstrates that the r-process elements in Ret II must have been synthesized early in its initial star-forming phase. We therefore constrain the delay time between the formation of the first stars in Ret II and the r-process nucleosynthesis to be less than 500 Myr. This measurement rules out an r-process source with a delay time of several Gyr or more, such as GW170817.
A Comparative Study of Deep Learning Models for Fracture and Pore Space Segmentation in Synthetic Fractured Digital Rocks
This study focuses on the comparative study of deep learning (DL) models for pore space and discrete fracture networks (DFNs) segmentation in synthetic fractured digital rocks, specifically targeting low-permeability rock formations, such as shale and tight sandstones. Accurate characterization of pore space and DFNs is critical for subsequent property analysis and fluid flow modeling. Four DL models, SegNet, U-Net, U-Net-wide, and nested U-Net (i.e., U-Net++), were trained, validated, and tested using synthetic datasets, including input and label image pairs with varying properties. The model performance was assessed regarding pixel-wise metrics, including the F1 score and pixel-wise difference maps. In addition, the physics-based metrics were considered for further analysis, including sample porosity and absolute permeability. Particularly, We first simulated the permeability of porous media containing only pore space and then simulated the permeability of porous media with DFNs added. The difference between these two values is used to quantify the connectivity of segmented DFNs, which is an important parameter for low-permeability rocks. The pixel-wise metrics showed that the nested U-Net model outperformed the rest of the DL models in pore space and DFNs segmentation, with the SegNet model exhibiting the second-best performance. Particularly, nested U-Net enhanced segmentation accuracy for challenging boundary pixels affected by partial volume effects. The U-Net-wide model achieved improved accuracy compared to the U-Net model, which indicated the influence of parameter numbers. Similarly, nested U-Net has the closest match to the ground truth of physics-based metrics, including the porosity of pore space and DFNs, and the permeability difference quantifying the connectivity of DFNs. The findings highlight the effectiveness of DL models, especially the U-Net++ model with nested architecture and redesigned skip connections, in accurately segmenting pore spaces and DFNs, which are crucial for pore-scale fluid flow and transport simulation in low-permeability rocks.
The solenoidal large intensity device (SoLID) for JLab 12 GeV
The solenoidal large intensity device (SoLID) is a new experimental apparatus planned for Hall A at the Thomas Jefferson National Accelerator Facility (JLab). SoLID will combine large angular and momentum acceptance with the capability to handle very high data rates at high luminosity. With a slate of approved high-impact physics experiments, SoLID will push JLab to a new limit at the QCD intensity frontier that will exploit the full potential of its 12 GeV electron beam. In this paper, we present an overview of the rich physics program that can be realized with SoLID, which encompasses the tomography of the nucleon in 3D momentum space from semi-inclusive deep inelastic scattering, expanding the phase space in the search for new physics and novel hadronic effects in parity-violating DIS, a precision measurement of J/ψ production at threshold that probes the gluon field and its contribution to the proton mass, tomography of the nucleon in combined coordinate and momentum space with deep exclusive reactions, and more. To meet the challenging requirements, the design of SoLID described here takes full advantage of recent progress in detector, data acquisition and computing technologies. In addition, we outline potential experiments beyond the currently approved program and discuss the physics that could be explored should upgrades of CEBAF become a reality in the future.
SoLID Program at JLab
An overview of the Solenoidal Large Intensity Device (SoLID) and its scientific program will be given in this talk. SoLID is a spectrometer/detector system proposed to exploit the full potential of the Jefferson Lab (JLab) 12 GeV energy upgrade. SoLID will push the limit of luminosity frontier in hadronic physics with its unique capability to handle very high rates with large acceptance under high luminosity (1037-39/cm2/s). A rich and vibrant scientific program has been developed for SoLID, including but not limited to the precision study of the 3d nucleon structure in both momentum space using Semi-Inclusive Deep Inelastic Scattering (SIDIS) and coordinate space using Deep Virtual Exclusive Reactions (DVER), probing physics beyond the Standard Model with Parity Violating Deep Inelastic Scattering (PVDIS), and investigating the gluonic field contribution to the proton structure and proton mass via J/¿ threshold production. The SoLID collaboration has developed a robust, low risk and flexible conceptual design, with a base line design capable of accomplishing its scientific goals and flexibility to adopt the cutting-edge technology. Detector subsystems have been tested with prototypes in realistic high luminosity conditions and are demonstrated to function well under extremely challenging environment to satisfy the requirements of planned experiments.
Exploring the Chemical Space of Linear Alkane Pyrolysis via Deep Potential GENerator
Not provided.
Deep Clustering based on Bi-Space Association Learning
Explore the source record for details and available documents.
Legacy Survey of Space and Time Data Preview 2: deep_coadd dataset type
We present Rubin Data Preview 2 (DP2), the second data preview from the NDF-DOE Vera C. Rubin Observatory. Data Preview 2 (DP2) comprises coadds, detection catalogs, and ancillary data products; and when fully released will also include single-epoch images and difference images. DP2 is derived from observations acquired by the LSST Science Camera (LSSTCam) on the Simonyi Survey Telescope at the Summit Facility on Cerro Pachón, Chile, primarily during the on-sky commissioning campaign between 2025-04-16 and 2025-09-21, supplemented by observations taken between 2025-10-25 and 2026-01-06 that overlap the commissioning footprint. The DP2 footprint comprises the Science Validation wide-area survey, five Deep Drilling Fields, and a number of targeted small-field regions, including Trifid and Lagoon, Prawn, M49, and New Horizons, all observed as part of the Rubin First Look campaign. Each field was imaged in up to six broad photometric bands, ugrizy, and coadded to produce deep imaging covering an estimated 3,000 deg2. The addition of single-visit-only areas expands the total DP2 footprint to an estimated 15,000 deg2, with coverage in at least one filter. The median per-visit PSF FWHM across the wide-area survey ranges from 1.17 arcsec in the z band to 1.26 arcsec in g and r bands. The deepest field, reaches estimated coadded 5σ depths of u=26 mag, g=26.8 mag, r=26.3 mag, i=26.1 mag, z=25.3 mag, y=23.9 mag. Based on a roughly five-month primary observing baseline and covering only part of the eventual LSST footprint, DP2's area, depth, and multiband coverage nonetheless support a broad range of early science investigations ahead of LSST Data Release This dataset is a subset of the full data release consisting of the deep_coadd dataset type. These are the combination of multiple processed, calibrated, and background- subtracted images, for a patch of sky, for each of the six filters. This release contains 925,460 datasets of this type.
Legacy Survey of Space and Time Data Preview 2: deep_coadd_input_summary dataset type
We present Rubin Data Preview 2 (DP2), the second data preview from the NDF-DOE Vera C. Rubin Observatory. Data Preview 2 (DP2) comprises coadds, detection catalogs, and ancillary data products; and when fully released will also include single-epoch images and difference images. DP2 is derived from observations acquired by the LSST Science Camera (LSSTCam) on the Simonyi Survey Telescope at the Summit Facility on Cerro Pachón, Chile, primarily during the on-sky commissioning campaign between 2025-04-16 and 2025-09-21, supplemented by observations taken between 2025-10-25 and 2026-01-06 that overlap the commissioning footprint. The DP2 footprint comprises the Science Validation wide-area survey, five Deep Drilling Fields, and a number of targeted small-field regions, including Trifid and Lagoon, Prawn, M49, and New Horizons, all observed as part of the Rubin First Look campaign. Each field was imaged in up to six broad photometric bands, ugrizy, and coadded to produce deep imaging covering an estimated 3,000 deg2. The addition of single-visit-only areas expands the total DP2 footprint to an estimated 15,000 deg2, with coverage in at least one filter. The median per-visit PSF FWHM across the wide-area survey ranges from 1.17 arcsec in the z band to 1.26 arcsec in g and r bands. The deepest field, reaches estimated coadded 5σ depths of u=26 mag, g=26.8 mag, r=26.3 mag, i=26.1 mag, z=25.3 mag, y=23.9 mag. Based on a roughly five-month primary observing baseline and covering only part of the eventual LSST footprint, DP2's area, depth, and multiband coverage nonetheless support a broad range of early science investigations ahead of LSST Data Release This dataset is a subset of the full data release consisting of the deep_coadd_input_summary dataset type. These are summary of the visit images that contribute to each deep coadd patch. This release contains 1 dataset of this type.
The Relationship Between Topography, Bedrock Weathering, and Water Storage Across a Sequence of Ridges and Valleys
Abstract Bedrock weathering regulates nutrient mobilization, water storage, and soil production. Relative to the mobile soil layer, little is known about the relationship between topography and bedrock weathering. Here, we identify a common pattern of weathering and water storage across a sequence of three ridges and valleys in the sedimentary Great Valley Sequence in Northern California that share a tectonic and climate history. Deep drilling, downhole logging, and characterization of chemistry and porosity reveal two weathering fronts. The shallower front is ∼7 m deep at the ridge of all three hillslopes, and marks the extent of pervasive fracturing and oxidation of pyrite and organic carbon. A deeper weathering front marks the extent of open fractures and discoloration. This front is 11 m deep under two ridges of similar ridge‐valley spacing, but 17.5 m deep under a ridge with nearly twice the ridge‐valley spacing. Hence, at ridge tops, the fraction of the hillslope relief that is weathered scales with hillslope length. In all three hillslopes, below this second weathering front, closed fractures and unweathered bedrock extend about one‐half the hilltop elevation above the adjacent channels. Neutron probe surveys reveal that seasonally dynamic moisture is stored to approximately the same depth as the shallow weathering front. Under the channels that bound our study hillslopes, the two weathering fronts coincide and occur within centimeters of the ground surface. Our findings provide evidence for feedbacks between erosion and weathering in mountainous landscapes that result in systematic subsurface structuring and water routing.
Investigating performance and variability of NIF ICF experiments with deep learning
The parameter space involved in designing an inertial confinement fusion shot at the National Ignition Facility (NIF) is massively multi-dimensional and the cost of a single shot makes a comprehensive set of sensitivity studies in the laboratory impractical. The use of machine learning to overcome these challenges has gained popularity and has had several successful applications by the scientific community. We extend on these efforts by training a neural network (NN) on information about the experimental design, engineering elements, and drive asymmetry to predict with uncertainty the neutron yield of an experiment. We find the measured and model predicted values are in good agreement, with an R 2 value of 0.91 for a randomly selected test dataset. Almost all the predicted 95% credible intervals contain the corresponding measured value for both training and test datasets. We identify correlations picked up by the NN between the shot design, yield, and variability and use them to motivate shot sensitivity studies. The first shot to exceed the Lawson-like ignition criteria (N210808) was conducted at the NIF and subsequent shots studied the design’s robustness. In a follow-up shot to N210808, our model predicts capsule quality to be the main performance degradation mechanism that prevented the shot from repeating previous performance levels. Shot N221204 was the first shot to exceed a target energy gain of 1. Our model predicts increased yield with reduced coast time for a N221204 study and greater variability for designs with lower peak powers at constant yield. The model’s fast prediction speed and uncertainty prediction are useful for identifying interesting design paths that could warrant further investigation with conventional simulations to search for robust high yield designs.
A reproducibility study of "Augmenting Genetic Algorithms with Deep Neural Networks for Exploring the Chemical Space"
Nigam et al. reported a genetic algorithm (GA) utilizing the SELFIES representation and also propose an adaptive, neural network-based penalty that is supposed to improve the diversity of the generated molecules. The main claims of the paper are that this GA outperforms other generative techniques (as measured by the penalized logP) and that a neural network-based adaptive penalty increases the diversity of the generated molecules. In this work, we investigated the reproducibility of their claims. Overall, we were able to reproduce comparable results using the SELFIES-based GA, but mostly by exploiting deficiencies of the (easily optimizable) fitness function (i.e., generating long, sulfur containing chains). In addition, we reproduce results showing that the discriminator can be used to bias the generation of molecules to ones that are similar to the reference set. Lastly, we attempted to quantify the evolution of the diversity, understand the influence of some hyperparameters, and propose improvements to the adaptive penalty.
A synthetic Roman Space Telescope High-Latitude Time-Domain Survey: supernovae in the deep field
ABSTRACT NASA will launch the Nancy Grace Roman Space Telescope (hereafter Roman) in the second half of this decade, which will allow for a generation-defining measurement of dark energy through multiple probes, including Type Ia supernovae (SNe Ia). To improve decisions on survey strategy, we have created the first simulations of realistic Roman images that include artificial SNe Ia injected as point sources in the images. Our analysis combines work done on Roman simulations for weak gravitational lensing studies as well as catalogue-level simulations of SN Ia samples. We have created a time series of images over 2 yr containing ∼1050 SNe Ia, covering a 1 deg2 subarea of a planned 5 deg2 deep survey. We have released these images publicly for community use along with input catalogues of all injected sources. We create secondary products from these images by generating coadded images and demonstrating recovery of transient sources using image subtraction. We perform first-use analyses on these images in order to measure galaxy detection efficiency, point source detection efficiency, and host-galaxy association biases. The simulated images can be found here at https://roman.ipac.caltech.edu/sims/SN_Survey_Image_sim.html.
Deep nonparametric estimation of operators between infinite dimensional spaces
Learning operators between infinitely dimensional spaces is an important learning task arising in machine learning, imaging science, mathematical modeling and simulations, etc. This paper studies the nonparametric estimation of Lipschitz operators using deep neural networks. Non-asymptotic upper bounds are derived for the generalization error of the empirical risk minimizer over a properly chosen network class. Under the assumption that the target operator exhibits a low dimensional structure, our error bounds decay as the training sample size increases, with an attractive fast rate depending on the intrinsic dimension in our estimation. Our assumptions cover most scenarios in real applications and our results give rise to fast rates by exploiting low dimensional structures of data in operator estimation. We also investigate the influence of network structures (e.g., network width, depth, and sparsity) on the generalization error of the neural network estimator and propose a general suggestion on the choice of network structures to maximize the learning efficiency quantitatively.