Search NASA⌕ Search

SEARCH · Search NASA

Results for “Dataset Curation”

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 289 records · Page 16
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Workshop Summary: Bridging the Gap Between Atmospheric Science and Grid Integration

The need for dedicated, accurate, expertly curated weather data is increasingly important as the share of variable renewable energy increases on the power system. Projections for futures with very high (50+% annual energy) shares of variable generation require ongoing assessment of data requirements from industry stakeholders in their power system operation and planning contexts. In March 2024, NREL organized a workshop entitled "Bridging the Gap Between Atmospheric Science and Grid Integration Workshop", which brought atmospheric scientists and power system experts together to refine the requirements of atmospheric datasets for grid integration, and to describe a holistic approach to creating new and regularly updated national scale wind datasets for power system planning and operations. The results of this workshop are being used to inform the near-term development and a longer-term strategy for DOE to produce relevant wind resource datasets and inform wider use of wind/solar/load data sets in power system planning. This presentation provides an overview of a preworkshop survey, an assessment of current state of the art of national-scale datasets for wind resource assessment and grid integration, insights on appropriate uses of the WTK-LED, power system perspectives on data needs, as well as recommended next steps as discussed in the workshop and how these steps support longer-term strategies.

17 WIND ENERGY↗
NASA NTRS2014

Aggregation Tool to Create Curated Data albums to Support Disaster Recovery and Response

Despite advances in science and technology of prediction and simulation of natural hazards, losses incurred due to natural disasters keep growing every year. Natural disasters cause more economic losses as compared to anthropogenic disasters. Economic losses due to natural hazards are estimated to be around $6-$10 billion dollars annually for the U.S. and this number keeps increasing every year. This increase has been attributed to population growth and migration to more hazard prone locations such as coasts. As this trend continues, in concert with shifts in weather patterns caused by climate change, it is anticipated that losses associated with natural disasters will keep growing substantially. One of challenges disaster response and recovery analysts face is to quickly find, access and utilize a vast variety of relevant geospatial data collected by different federal agencies such as DoD, NASA, NOAA, EPA, USGS etc. Some examples of these data sets include high spatio-temporal resolution multi/hyperspectral satellite imagery, model prediction outputs from weather models, latest radar scans, measurements from an array of sensor networks such as Integrated Ocean Observing System etc. More often analysts may be familiar with limited, but specific datasets and are often unaware of or unfamiliar with a large quantity of other useful resources. Finding airborne or satellite data useful to a natural disaster event often requires a time consuming search through web pages and data archives. Additional information related to damages, deaths, and injuries requires extensive online searches for news reports and official report summaries. An analyst must also sift through vast amounts of potentially useful digital information captured by the general public such as geo-tagged photos, videos and real time damage updates within twitter feeds. Collecting and aggregating these information fragments can provide useful information in assessing damage in real time and help direct recovery efforts. The search process for the analyst could be made much more efficient and productive if a tool could go beyond a typical search engine and provide not just links to web sites but actual links to specific data relevant to the natural disaster, parse unstructured reports for useful information nuggets, as well as gather other related reports, summaries, news stories, and images. This presentation will describe a semantic aggregation tool developed to address similar problem for Earth Science researchers. This tool provides automated curation, and creates "Data Albums" to support case studies. The generated "Data Albums" are compiled collections of information related to a specific science topic or event, containing links to relevant data files (granules) from different instruments; tools and services for visualization and analysis; information about the event contained in news reports, and images or videos to supplement research analysis. An ontology-based relevancy-ranking algorithm drives the curation of relevant data sets for a given event. This tool is now being used to generate a catalog of Hurricane Case Studies at Global Hydrology Resource Center (GHRC), one of NASA's Distribute Active Archive Centers. Another instance of the Data Albums tool is currently being created in collaboration with NASA/MSFC's SPoRT Center, which conducts research on unique NASA products and capabilities that can be transitioned to the operational community to solve forecast problems. This new instance focuses on severe weather to support SPoRT researchers in their model evaluation studies

Ramachandran, Rahul↗
NASA NTRSDate not supplied

Blast from the Past: ASDC Curation for NASA Suborbital Legacy Missions to Promote Data Discovery and Accessibility

NASA has an extensive history of conducting suborbital field campaigns to further advances in atmospheric sciences. Beginning with the Chemical Instrument Test and Evaluation (CITE) conducted in 1983-1984, NASA has completed many suborbital campaigns over the past three decades. Since the early 2010s, suborbital missions are typically assigned to a NASA Distributed Active Archive Center (DAAC) prior to the mission for long-term archival and distribution. Efforts are being made by NASA’s Earth Science Data and Information System (ESDIS) Project and the Airborne Data Management Group (ADMG) to assign legacy missions to DAACs for permanent archival and distribution, so that these valuable datasets remain to be available to the scientific community. NASA’s Atmospheric Science Data Center (ASDC) has been named the assigned DAAC for nearly 20 atmospheric composition legacy missions, including missions conducted as part of the Global Tropospheric Experiment (GTE) and expects to be named the assigned DAAC for more of these missions over the next few years. The primary goal of the ASDC is to provide access to the datasets as they are currently formatted to the broad user community and enhance their findability and accessibility. However, data reporting standards have evolved significantly since 1983 and the datasets span a wide variety of file formats, including text, Ames, GTE, and ICARTT (International Consortium for Atmospheric Research on Transport and Transformation), and the amount of metadata and relevant information included in the files also varies greatly and can not be readily extracted without subject matter knowledge. This has caused challenges for the ASDC’s suborbital metadata extraction pipeline in ensuring that accurate and necessary metadata is being provided for the missions by all the ASDC’s existing search mechanisms. To make the data more findable and accessible, the ASDC has begun researching ways to further enhance the datasets, including distributing value-added products (i.e. consistent file format such as ICARTT or netCDF), adding standard names from the ESDIS Standards Coordination Office (ESCO)-approved Atmospheric Composition Variable Standard Names Convention (ACVSNC), and creating outreach materials such as ArcGIS StoryMaps, User Guides, and Micro Articles, providing overviews of the missions and what type of data was collected during the missions. These efforts also help support NASA’s Open-Source Science by enhancing the FAIRness of the legacy data products. This presentation will review the ASDC’s ongoing efforts, progress made, and future plans for legacy missions.

Megan Buzanowicz↗
DOE OSTI2026

A reproducible study design for the MIMIC-IV in-hospital mortality task

Open, tabular electronic health record (EHR) datasets such as MIMIC-III and MIMIC-IV have become critical resources for developing machine learning (ML) models addressing clinical prediction tasks, including hospital readmission, length of stay, and in-hospital mortality (IHM). While MIMIC-III has benefited from well-established preprocessing pipelines and standardized feature sets, MIMIC-IV remains comparatively challenging to work with because there are no standardized benchmarks to support reproducibility and comparability across studies. To address this limitation, we present a rigorously curated MIMIC-IV custom feature set optimized for IHM prediction, constructed through a reproducible preprocessing pipeline and feature selection strategy.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

OrthoPhyl—streamlining large-scale, orthology-based phylogenomic studies of bacteria at broad evolutionary scales

Abstract There are a staggering number of publicly available bacterial genome sequences (at writing, 2.0 million assemblies in NCBI's GenBank alone), and the deposition rate continues to increase. This wealth of data begs for phylogenetic analyses to place these sequences within an evolutionary context. A phylogenetic placement not only aids in taxonomic classification but informs the evolution of novel phenotypes, targets of selection, and horizontal gene transfer. Building trees from multi-gene codon alignments is a laborious task that requires bioinformatic expertise, rigorous curation of orthologs, and heavy computation. Compounding the problem is the lack of tools that can streamline these processes for building trees from large-scale genomic data. Here we present OrthoPhyl, which takes bacterial genome assemblies and reconstructs trees from whole genome codon alignments. The analysis pipeline can analyze an arbitrarily large number of input genomes (>1200 tested here) by identifying a diversity-spanning subset of assemblies and using these genomes to build gene models to infer orthologs in the full dataset. To illustrate the versatility of OrthoPhyl, we show three use cases: E. coli/Shigella, Brucella/Ochrobactrum and the order Rickettsiales. We compare trees generated with OrthoPhyl to trees generated with kSNP3 and GToTree along with published trees using alternative methods. We show that OrthoPhyl trees are consistent with other methods while incorporating more data, allowing for greater numbers of input genomes, and more flexibility of analysis.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Seismic Signal Detection on International Monitoring System 3-Component Stations using PhaseNet

In this report we discuss training a deep learning seismic signal detection model on 3-component stations from the International Monitoring System (IMS) using the PhaseNet architecture. Using 14 years of associated signals from the International Data Centre’s (IDC) Late Event Bulletin (LEB), we auto-curated training data consisting of signal windows containing associated arrivals, and noise windows that contain no LEB-associated signals. We trained several models using different waveform window durations (30 seconds and 100 seconds), with and without bandpass filtering. We evaluated the effectiveness of our models using associated signals from the Unconstrained Global Event Bulletin (UGEB) and found that several of our models outperformed the signal detections from the IDC’s Selected Event List 3 (SEL3) arrival table. The SEL3 bulletin evaluated on the UGEB dataset with 100-second waveform windows registered a precision and recall of .15 and .48, respectively, versus .19 and .59 for our filtered-data model. For the 30-second waveform window dataset, the SEL3 bulletin achieved a precision and recall of .31 and .47, respectively, versus .32 and .60 for our filtered-data model. Finally, our models detected signals from all source-to-receiver distances, suggesting it is feasible to use a single PhaseNet model for the IMS network.

58 GEOSCIENCES↗
DOE OSTI2025

Large language model-driven database for thermoelectric materials

Thermoelectric materials have the ability to convert waste heat into electricity, offering a valuable solution for energy harvesting. However, their widespread use is hindered by low conversion efficiency, the reliance on expensive rare earth elements, and the environmental and regulatory concerns associated with lead-based materials. A fast and cost-effective way to identify highly efficient thermoelectric materials is through data-driven methods. These approaches rely on robust and comprehensive datasets to train models. Although there are several databases on thermoelectric materials, there is still a need to collect and integrate experimental data from peer-reviewed research articles to capture diverse compositions and properties of materials. Here, in this work, we developed a comprehensive database of 7,123 thermoelectric compounds, containing key information such as chemical composition, structural detail, seebeck coefficient, electrical and thermal conductivity, power factor, and figure of merit (ZT). We used the GPTArticleExtractor workflow, powered by large language models (LLM), to extract and curate data automatically from the scientific literature published in Elsevier journals. This process enabled the creation of a structured database that addresses the challenges of manual data collection. The open access database could stimulate data-driven research and advance thermoelectric material analysis and discovery.

Database↗
DOE OSTI2021

Tripal, a community update after 10 years of supporting open source, standards-based genetic, genomic and breeding databases

Abstract Online, open access databases for biological knowledge serve as central repositories for research communities to store, find and analyze integrated, multi-disciplinary datasets. With increasing volumes, complexity and the need to integrate genomic, transcriptomic, metabolomic, proteomic, phenomic and environmental data, community databases face tremendous challenges in ongoing maintenance, expansion and upgrades. A common infrastructure framework using community standards shared by many databases can reduce development burden, provide interoperability, ensure use of common standards and support long-term sustainability. Tripal is a mature, open source platform built to meet this need. With ongoing improvement since its first release in 2009, Tripal provides full functionality for searching, browsing, loading and curating numerous types of data and is a primary technology powering at least 31 publicly available databases spanning plants, animals and human data, primarily storing genomics, genetics and breeding data. Tripal software development is managed by a shared, inclusive governance structure including both project management and advisory teams. Here, we report on the most important and innovative aspects of Tripal after 11 years development, including integration of diverse types of biological data, successful collaborative projects across member databases, and support for implementing FAIR principles.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Developing a SARS-CoV-2 main protease binding prediction random forest model for drug repurposing for COVID-19 treatment

The coronavirus disease 2019 (COVID-19) global pandemic resulted in millions of people becoming infected with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus and close to seven million deaths worldwide. It is essential to further explore and design effective COVID-19 treatment drugs that target the main protease of SARS-CoV-2, a major target for COVID-19 drugs. In this study, machine learning was applied for predicting the SARS-CoV-2 main protease binding of Food and Drug Administration (FDA)-approved drugs to assist in the identification of potential repurposing candidates for COVID-19 treatment. Ligands bound to the SARS-CoV-2 main protease in the Protein Data Bank and compounds experimentally tested in SARS-CoV-2 main protease binding assays in the literature were curated. These chemicals were divided into training (516 chemicals) and testing (360 chemicals) data sets. To identify SARS-CoV-2 main protease binders as potential candidates for repurposing to treat COVID-19, 1188 FDA-approved drugs from the Liver Toxicity Knowledge Base were obtained. A random forest algorithm was used for constructing predictive models based on molecular descriptors calculated using Mold2 software. Model performance was evaluated using 100 iterations of fivefold cross-validations which resulted in 78.8% balanced accuracy. The random forest model that was constructed from the whole training dataset was used to predict SARS-CoV-2 main protease binding on the testing set and the FDA-approved drugs. Model applicability domain and prediction confidence on drugs predicted as the main protease binders discovered 10 FDA-approved drugs as potential candidates for repurposing to treat COVID-19. Our results demonstrate that machine learning is an efficient method for drug repurposing and, thus, may accelerate drug development targeting SARS-CoV-2.

Research & Experimental Medicine↗
DOE OSTI2026

Meta-virus resource (MetaVR): expanding the frontiers of viral diversity with 24 million uncultivated virus genomes

Viruses are ubiquitous in all environments and impact host metabolism, evolution, and ecology, although our knowledge of their biodiversity is still extremely limited. Viral diversity from genomic and metagenomic datasets has led to an explosion of uncultivated virus genomes (UViGs) and the development of specialized databases to catalog this viral diversity, though many lack comprehensive integration. Here, we introduce meta-virus resource (MetaVR), the successor of the IMG/VR database, designed to overcome previous limitations such as large-scale querying and programmatic access. Drawing on the increase of publicly available genomes and metagenomes, MetaVR significantly expands viral diversity, now comprising 24,435,662 UViGs, a 57.6% increase from its predecessor, organized into over 12 million viral operational taxonomic units. Key enhancements include the integration of curated eukaryotic host information, the integration of protein clusters and predicted structures for comparative studies, and an API for programmatic data access. Furthermore, MetaVR features an updated taxonomic framework based on ICTV release 39, assignment to Baltimore classes, and enhanced host assignment through novel computational tools like iPHoP. These advancements position MetaVR as a unique resource for exploring viral diversity, evolution, and host interactions across diverse environments. MetaVR can be freely accessed at https://www.meta-virome.org/.

Fiamenghi, Mateus B↗
DOE OSTI2022

EXSCLAIM!

Due to recent improvements in image resolution and acquisition speed, materials microscopy is experiencing an explosion of published imaging data. The standard publication format, while sufficient for traditional data ingestion scenarios where a select number of images can be critically examined and curated manually, is not conducive tolarge-scale data aggregation or analysis, hindering data sharing and reuse. Most images in publications are presented as components of a larger figure with their explicit context buried in the main body or caption text, so even if aggregated, collections of images with weak or no digitized contextual labels have limited value. To solve the problem of curating labeled microscopy data from literature, we introduce the EXSCLAIM! Python toolkit for the automatic EXtraction, Separation, and Caption-based natural Language Annotation of IMages from scientific literature. The software is implemented through a three part pipeline: the JournalScraper, which searches the web and downloads figures and captions based on a user provided query, the CaptionDistributor, which separates caption text based on the subfigure each portion of the caption refers to, and the FigueSeparator, which separates figures into component subfigures and extracts other visual information. Also included is a Django user interface for exploring the resulting dataset.

CHAN, MARIA↗
NASA NTRSDate not supplied

Spaceflight Biospecimen Sharing in Support of Science Discovery and Exploration

For decades, NASA and international partners have flown non-human biological experiments in space to understand the effects of spaceflight and address potential biological hazards. Sending organisms into space is a costly endeavor which makes space-flown biological specimens a valuable resource. To enable maximum scientific return, samples not required by the Principal Investigators are harvested and collected mostly by NASA’s Space Biology Biospecimen Sharing Program. These specimens are collected according to well-established SOPs that maintain quality and integrity. The specimens are then preserved, archived, and made available to the international scientific community through NASA’s Institutional Scientific Collection (ISC) at Ames Research Center (ARC). The ISC-ARC biospecimens and descriptive metadata are findable and accessible for request through the Life Sciences Data Archive (LSDA). The NASA ISC-ARC currently stores over 32,000 specimens from Shuttle, International Space Station, and ground-based investigations (spaceflight analog experiments involving either hindlimb unloading, centrifugation, or partial weight-bearing study designs). Tissues are predominantly from mice and rats, though samples are also available from bacteria and quail. The specimens include tissues from many physiological systems including musculoskeletal, neurosensory, reproductive, respiratory, circulatory, and digestive. Tissues are stored at -80°C, -20°C, +4°C, or ambient and preserved in various fixatives. Descriptive metadata is available for all samples. Historically, these tissues have been used for a wide range of analyses, including histology, genomics, and transcriptomics. Plans are underway to expand the ISC-ARC beyond the mostly-rodent contents, to include a space-relevant microbial culture collection including bacteria, fungi, and yeast. This expansion of the ISC-ARC will now involve identifying and standardizing best practices for microbial curations. To ensure safe long-term storage of microbial isolates, a microbiology laboratory will be dedicated for identification, cell culture, and lyophilization. Awarding of tissue to public science investigators has resulted in 33 publications since 2011, with 48 requests being submitted since 2016. Of note, NASA GeneLab has been awarded ISC-ARC biospecimens in the past few years. GeneLab processes the biospecimens to generate various levels of ‘omics’ data, which are published on GeneLab’s open access online platform for bioinformatics analysis and visualization. This has helped a systems biology community grow around the processed-biospecimens’ datasets, resulting in many new publications and insights. Websites: https://www.nasa.gov/ames/research/space-biosciences/isc-bsp ; https://lsda.jsc.nasa.gov/Biospecimen

Ryan T. Scott↗
NASA NTRS2022

Introducing and Evaluating the Climate Hazards Center IMERG with Stations (CHIMES) Timely Station-Enhanced Integrated Multisatellite Retrievals for Global Precipitation Measurement

As human exposure to hydroclimatic extremes increase and the number of in situ precipitation observations declines, precipitation estimates, such as those provided by the Integrated Multisatellite Retrievals for Global Precipitation Measurement (GPM) (IMERG) mission, provide a critical source of information. Here, we present a new gauge-enhanced dataset [the Climate Hazards Center IMERG with Stations (CHIMES)] designed to support global crop and hydrologic modeling and monitoring. CHIMES enhances the IMERG Late Run product using an updated Climate Hazards Center (CHC) high-resolution climatology (CHPclim) and low-latency rain gauge observations. CHPclim differs from other products because it incorporates long-term averages of satellite precipitation, which increases CHPclim ’s fidelity in data-sparse areas with complex terrain. This fidelity translates into performance increases in unbiased IMERGlate data, which we refer to as CHIME. This is augmented with gauge observations to produce CHIMES. The CHC’s curated rain gauge archive contains valuable contributions from many countries. There are two versions of CHIMES: preliminary and final. The final product has more copious and better-curated station data. Every pentad and month, bias-adjusted IMERGlate fields are combined with gauge observations to create pentadal and monthly CHIMESprelim and CHIMESfinal. Comparisons with pentadal, high-quality gridded station data show that IMERG late performs well (r = 0.75), but has some systematic biases which can be reduced. Monthly cross-validation results indicate that unbiasing increases the variance explained from 50% to 63% and decreases the mean absolute error from 48 to 39 mm month −1. Gauge enhancement then increases the variance explained to 75%, reducing the mean absolute error to 27 mm month −1.

Chris C funk↗
DOE OSTI2024

Machine learning analysis of RB-TnSeq fitness data predicts functional gene modules in Pseudomonas putida KT2440

ABSTRACT There is growing interest in engineering Pseudomonas putida KT2440 as a microbial chassis for the conversion of renewable and waste-based feedstocks, and metabolic engineering of P. putida relies on the understanding of the functional relationships between genes. In this work, independent component analysis (ICA) was applied to a compendium of existing fitness data from randomly barcoded transposon insertion sequencing (RB-TnSeq) of P. putida KT2440 grown in 179 unique experimental conditions. ICA identified 84 independent groups of genes, which we call fModules (“functional modules”), where gene members displayed shared functional influence in a specific cellular process. This machine learning-based approach both successfully recapitulated previously characterized functional relationships and established hitherto unknown associations between genes. Selected gene members from fModules for hydroxycinnamate metabolism and stress resistance, acetyl coenzyme A assimilation, and nitrogen metabolism were validated with engineered mutants of P. putida . Additionally, functional gene clusters from ICA of RB-TnSeq data sets were compared with regulatory gene clusters from prior ICA of RNAseq data sets to draw connections between gene regulation and function. Because ICA profiles the functional role of several distinct gene networks simultaneously, it can reduce the time required to annotate gene function relative to manual curation of RB-TnSeq data sets. IMPORTANCE This study demonstrates a rapid, automated approach for elucidating functional modules within complex genetic networks. While Pseudomonas putida randomly barcoded transposon insertion sequencing data were used as a proof of concept, this approach is applicable to any organism with existing functional genomics data sets and may serve as a useful tool for many valuable applications, such as guiding metabolic engineering efforts in other microbes or understanding functional relationships between virulence-associated genes in pathogenic microbes. Furthermore, this work demonstrates that comparison of data obtained from independent component analysis of transcriptomics and gene fitness datasets can elucidate regulatory-functional relationships between genes, which may have utility in a variety of applications, such as metabolic modeling, strain engineering, or identification of antimicrobial drug targets.

09 BIOMASS FUELS↗
DOE OSTI2021

A Comprehensive Urine Proteome Database Generated From Patients With Various Renal Conditions and Prostate Cancer

Urine proteins can serve as viable biomarkers for diagnosing and monitoring various diseases. A comprehensive urine proteome database, generated from a variety of urine samples with different disease conditions, can serve as a reference resource for facilitating discovery of potential urine protein biomarkers. Herein, we present a urine proteome database generated from multiple datasets using 2D LC-MS/MS proteome profiling of urine samples from healthy individuals (HI), renal transplant patients with acute rejection (AR) and stable graft (STA), patients with non-specific proteinuria (NS), and patients with prostate cancer (PC). A total of ~28,000 unique peptides spanning ~2,200 unique proteins were identified with a false discovery rate of <0.5% at the protein level. Over one third of the annotated proteins were plasma membrane proteins and another one third were extracellular proteins according to gene ontology analysis. Ingenuity Pathway Analysis of these proteins revealed 349 potential biomarkers in the literature-curated database. Forty-three percentage of all known cluster of differentiation (CD) proteins were identified in the various human urine samples. Interestingly, following comparisons with five recently published urine proteome profiling studies, which applied similar approaches, there are still ~400 proteins which are unique to this current study. These may represent potential disease-associated proteins. Among them, several proteins such as serpin B3, renin receptor, and periostin have been reported as pathological markers for renal failure and prostate cancer, respectively. Taken together, our data should provide valuable information for future discovery and validation studies of urine protein biomarkers for various diseases.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Improving Cyber Situational Understanding

Effective cybersecurity operations require the ability to analyze large amounts of information to assess security risks and formulate defensive strategies against adversaries. This has become more complex in recent years as the sprawl and interconnectivity of devices grows through implementation of virtualization, cloud computing, and Internet of Things (IoT). The amount of data and analysis required for effective cybersecurity command and control decisions far exceeds humans’ capacity to perform manually. We characterize the analysis problem as cyber situational understanding. The research presented to improve cyber situational understanding focuses on vulnerability analysis and threat intelligence. Regarding vulnerabilities, entities must analyze and plan work for between thousands and tens of thousands of software vulnerabilities annually. Entities heavily use network firewalls to limit vulnerability exposure. As a result, some of these vulnerabilities permit exposure to adversarial exploitation, whereas others are inaccessible and therefore present negligible risk of exploitation. Distinguishing between high and low risk software vulnerabilities requires a deep understanding of the vulnerability, network firewall protection, and characteristics of the targeted device. This problem is solved by extracting network service features from vulnerability data features using both machine-learning and natural language processing. Then, the network firewall topology is parsed to determine which vulnerabilities are reachable by adversaries. Ultimately, a state-based safety analysis ascertains which vulnerabilities are unsafe. A related vulnerability analysis problem occurs in cybersecurity operations when associating an entity’s hardware and software assets to public vulnerability databases. Assets often reveal hardware and software through installation artifacts and network service identification, and entities store these artifacts in inventory databases. However, software and hardware vendors apply a standard Common Platform Enumeration (CPE) naming convention when publicly reporting vulnerabilities. Associating these two datasets often requires many hours to days of manual inspection. The proposed solution automates the mapping approach of human analysts using fuzzy matching techniques, natural language processing, and, ultimately, machine learning to present a small set of recommendations for mapping the two datasets. The result significantly reduces human analysis time and reduces the occurrence of false positives in vulnerability notifications. Finally, cyber threat intelligence (CTI) requires associating cyber observable artifacts, such as IP addresses, URIs, and file hashes, with cyber threat tactics, techniques, and procedures. Unfortunately, most CTI data is compartmentalized across multiple organizations and cannot be shared due to the legal and reputational risk with cyber threat being associated with the entity. The approach to solving this problem inovlves using a distributed ledger with anonymous token spending and authentication. This allows a consortium of semi-trusted entities to share the workload of curating CTI for a threat sharing community’s cooperative benefit.

Huff, Philip↗
DOE OSTI2026

Identifying Anomalous DESI Galaxy Spectra with a Variational Autoencoder

The tens of millions of spectra being captured by the Dark Energy Spectroscopic Instrument (DESI) provide tremendous discovery potential. In this work we show how Machine Learning, in particular Variational Autoencoders (VAE), can detect anomalies in a sample of approximately 200,000 DESI spectra comprising galaxies, quasars and stars. We demonstrate that the VAE can compress the dimensionality of a spectrum by a factor of 100, while still retaining enough information to accurately reconstruct spectral features. We then detect anomalous spectra as those with high reconstruction error and those which are isolated in the VAE latent representation. The anomalies identified fall into two categories: spectra with artefacts and spectra with unique physical features. Awareness of the former can help to improve the DESI spectroscopic pipeline; whilst the latter can lead to the identification of new and unusual objects. To further curate the list of outliers, we use the Astronomaly package which employs Active Learning to provide personalised outlier recommendations for visual inspection. In this work we also explore the VAE latent space, finding that different object classes and subclasses are separated despite being unlabelled. We demonstrate the interpretability of this latent space by identifying tracks within it that correspond to various spectral characteristics. For example, we find tracks that correspond to increasing star formation and increase in broad emission lines along the Balmer series. In upcoming work we hope to apply the methods presented here to search for both systematics and astrophysically interesting objects in much larger datasets of DESI spectra.

Nicolaou, C. [University Coll. London] (ORCID:0000↗
NASA NTRS2020

Amino Acid Encoding for Deep Learning Applications

Background: The number of applications of deep learning algorithms in bioinformatics is increasing as they usually achieve superior performance over classical approaches, especially, when bigger training datasets are available. In deep learning applications, discrete data, e.g. words or n-grams in language, or amino acids or nucleotides in bioinformatics, are generally represented as a continuous vector through an embedding matrix. Recently, learning this embedding matrix directly from the data as part of the continuous iteration of the model to optimize the target prediction – a process called ‘end-to-end learning’ – has led to state-ofthe-art results in many fields. Although usage of embeddings is well described in the bioinformatics literature, the potential of end-to-end learning for single amino acids, as compared to more classical manually-curated encoding strategies, has not been systematically addressed. To this end, we compared classical encoding matrices, namely one-hot, VHSE8 and BLOSUM62, to end-to-end learning of amino acid embeddings for two different prediction tasks using three widely used architectures, namely recurrent neural networks (RNN), convolutional neural networks (CNN), and the hybrid CNN-RNN. Results: By using different deep learning architectures, we show that end-to-end learning is on par with classical encodings for embeddings of the same dimension even when limited training data is available, and might allow for a reduction in the embedding dimension without performance loss, which is critical when deploying the models to devices with limited computational capacities. We found that the embedding dimension is a major factor in controlling the model performance. Surprisingly, we observed that deep learning models are capable of learning from random vectors of appropriate dimension. Conclusion: Our study shows that end-to-end learning is a flexible and powerful method for amino acid encoding. Further, due to the flexibility of deep learning systems, amino acid encoding schemes should be benchmarked against random vectors of the same dimension to disentangle the information content provided by the encoding scheme from the distinguishability effect provided by the scheme.

Hesham ElAbd↗