Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “human domain”

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 793 records · Page 44
All sourcesNASA NTRSDOE OSTI
NASA NTRS1999

Research Institute for Advanced Computer Science: Annual Report October 1998 through September 1999

The Research Institute for Advanced Computer Science (RIACS) carries out basic research and technology development in computer science, in support of the National Aeronautics and Space Administration's missions. RIACS is located at the NASA Ames Research Center (ARC). It currently operates under a multiple year grant/cooperative agreement that began on October 1, 1997 and is up for renewal in the year 2002. ARC has been designated NASA's Center of Excellence in Information Technology. In this capacity, ARC is charged with the responsibility to build an Information Technology Research Program that is preeminent within NASA. RIACS serves as a bridge between NASA ARC and the academic community, and RIACS scientists and visitors work in close collaboration with NASA scientists. RIACS has the additional goal of broadening the base of researchers in these areas of importance to the nation's space and aeronautics enterprises. RIACS research focuses on the three cornerstones of information technology research necessary to meet the future challenges of NASA missions: (1) Automated Reasoning for Autonomous Systems. Techniques are being developed enabling spacecraft that will be self-guiding and self-correcting to the extent that they will require little or no human intervention. Such craft will be equipped to independently solve problems as they arise, and fulfill their missions with minimum direction from Earth. (2) Human-Centered Computing. Many NASA missions require synergy between humans and computers, with sophisticated computational aids amplifying human cognitive and perceptual abilities; (3) High Performance Computing and Networking Advances in the performance of computing and networking continue to have major impact on a variety of NASA endeavors, ranging from modeling and simulation to data analysis of large datasets to collaborative engineering, planning and execution. In addition, RIACS collaborates with NASA scientists to apply information technology research to a variety of NASA application domains. RIACS also engages in other activities, such as workshops, seminars, and visiting scientist programs, designed to encourage and facilitate collaboration between the university and NASA information technology research communities.

Leiner, Barry M.↗
NASA NTRS2000

Research Institute for Advanced Computer Science

The Research Institute for Advanced Computer Science (RIACS) carries out basic research and technology development in computer science, in support of the National Aeronautics and Space Administration's missions. RIACS is located at the NASA Ames Research Center. It currently operates under a multiple year grant/cooperative agreement that began on October 1, 1997 and is up for renewal in the year 2002. Ames has been designated NASA's Center of Excellence in Information Technology. In this capacity, Ames is charged with the responsibility to build an Information Technology Research Program that is preeminent within NASA. RIACS serves as a bridge between NASA Ames and the academic community, and RIACS scientists and visitors work in close collaboration with NASA scientists. RIACS has the additional goal of broadening the base of researchers in these areas of importance to the nation's space and aeronautics enterprises. RIACS research focuses on the three cornerstones of information technology research necessary to meet the future challenges of NASA missions: (1) Automated Reasoning for Autonomous Systems. Techniques are being developed enabling spacecraft that will be self-guiding and self-correcting to the extent that they will require little or no human intervention. Such craft will be equipped to independently solve problems as they arise, and fulfill their missions with minimum direction from Earth; (2) Human-Centered Computing. Many NASA missions require synergy between humans and computers, with sophisticated computational aids amplifying human cognitive and perceptual abilities; (3) High Performance Computing and Networking. Advances in the performance of computing and networking continue to have major impact on a variety of NASA endeavors, ranging from modeling and simulation to data analysis of large datasets to collaborative engineering, planning and execution. In addition, RIACS collaborates with NASA scientists to apply information technology research to a variety of NASA application domains. RIACS also engages in other activities, such as workshops, seminars, and visiting scientist programs, designed to encourage and facilitate collaboration between the university and NASA information technology research communities.

Gross, Anthony R.↗
NASA NTRS2019

Cluster Node Computing for Target Generation Systems in Aircraft Simulations

Target generation systems provide the positions of aircraft in air traffic simulations. As the scope of the simulation domains expand, there is a need to develop systems that can provide position reports for thousands of aircraft simultaneously and at update rates that support out-the-window visualization. This paper discusses the motivation and reasoning behind investigating development of a next generation target generator through distributed computing using clustered node processing and how a target generation system benefit future research that utilizes human-in-the-loop simulations.

air traffic simulations↗
NASA NTRS2010

Distribution of Chromosome Breakpoints in Human Epithelial Cells Exposed to Low- and High-LET Radiation

Low-and high-LET radiations produced distinct breakpoint distributions. The difference of the breakpoint distributions between low-and high-LET only appeared in break ends involved in interchromosome exchanges. The breakpoint distributions for break ends participating in intrachromosome exchanges were similar. Gene-rich regions do not necessarily have more chromosome breaks. High-LET appeared to produce long live (data not shown) or longer live breaks that can migrate a longer distance before rejoining with other breaks. Domains occupied by different segments of the chromosomes may be responsible for the breakpoint distribution. The dose responses for interchromosomal exchanges were linear in all four exposures. However, the dose response for intrachromosomal exchanges were none linear. Increasing dose of high dose rate exposure (Fe-ions or -rays) increase the fraction of cells with intrachromosome aberrations, whereas increasing dose of low dose rate exposure (neutrons or -rays) does not affect the fraction of cells with intrachromosome aberrations.

Hada, Megumi↗
NASA NTRSDate not supplied

Realizing a Spacesuit Compatible Augmented Reality System to Meet the Work Needs of Future Human Spaceflight Exploration

In this paper, we identify the broad range of challenges associated with creating an augmented reality for future spaceflight missions. The challenges were framed within a sociotechnical framework to explore how an imagined augmented reality system might one day help crew cope with future exploration mission demands. Key challenges include an augmented reality system cannot simply be designed to help a single suited astronaut and must enable the broader mission team. There exists a host of optical, compute, and control technical challenges that must be simultaneously addressed while the work domain itself is in development. There exist limited methods to help cope with the envisioned world problem; however, we emphasize a few strategies to articulate work design assumptions and host this detail in a transparent fashion. Finally, we note the broader organizational constraints that shape the creation of such novel technology for spaceflight. By addressing these broad range of challenges directly, early in the design and development process, key trade-offs can be considered and fundamental concerns within the community can be addressed to yield desired mission outcomes.

Matthew J. Miller↗
NASA NTRS2004

Structural Basis for the Catalytic Activity of Human Serine/Threonine Protein Phosphatase type 5 (PP5)

Serine/threonine protein phosphatase-5 (PP5) is a member of the PPP-gene family of protein phosphatases that is widely expressed in mammalian tissues and is highly conserved among eukaryotes. PP5 associates with several proteins that affect signal transduction networks, including the glucocorticoid receptor (GR)-heat shock protein-90 (Hsp90)-heterocomplex, the CDC16 and CDC27 subunits of the anaphase-promoting complex, elF2alpha kinase, the A subunit of PP2A, the G12-alpha / G13-alpha subunits of heterotrimeric G proteins and DNA-PK. The catalytic domain of PP5 (PP5c) shares 35-45% sequence identity with the catalytic domains of other PPP-phosphatases, including protein phosphatase-1 (PP1), -2A (PP2A), -2B / calcineurin (PP2B), -4 (PP4), -6 (PP6), and -7 (PP7). Like PP1, PP2A and PP4, PP5 is also sensitive to inhibition by okadaic acid, microcystin, cantharidin, tautomycin, and calyculin A. Here we report the crystal structure of the PP5 catalytic domain (PP5c) at a resolution of 1.6 angstroms. From this structure we propose a mechanism for PP5-mediated hydrolysis of phosphoprotein substrates, which requires the precise positioning of two metal ions within a conserved Asp(sup 271)-M(sub 1):M(sub 2)-W(sup 1)-His(sup 304)-Asp(sup 274) catalytic motif. The structure of PP5c provides a possible structural basis for explaining the exceptional catalytic proficiency of protein phosphatases, which are among the most powerful known catalysts. Resolution of the entire C-terminus revealed a novel subdomain, and the structure of the PP5c should also aid development of type-specific inhibitors.

Swingle, Mark R.↗
DOE OSTI2020

A 3.4-Å cryo-electron microscopy structure of the human coronavirus spike trimer computationally derived from vitrified NL63 virus particles

Human coronavirus NL63 (HCoV-NL63) is an enveloped pathogen of the family Coronaviridae that spreads worldwide and causes up to 10% of all annual respiratory diseases. HCoV-NL63 is typically associated with mild upper respiratory symptoms in children, elderly and immunocompromised individuals. It has also been shown to cause severe lower respiratory illness. NL63 shares ACE2 as a receptor for viral entry with SARS-CoV-1 and SARS-CoV-2. Here we present the in situ structure of HCoV-NL63 spike (S) trimer at 3.4-Å resolution by single-particle cryo-EM imaging of vitrified virions without chemical fixative. It is structurally homologous to that obtained previously from the biochemically purified ectodomain of HCoV-NL63 S trimer, which displays a 3-fold symmetric trimer in a single conformation. In addition to previously proposed and observed glycosylation sites, our map shows density at other sites, as well as different glycan structures. The domain arrangement within a protomer is strikingly different from that of the SARS-CoV-2 S and may explain their different requirements for activating binding to the receptor. Overall, this structure provides the basis for future studies of spike proteins with receptors, antibodies, or drugs, in the native state of the coronavirus particles.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

LLMs for Mfg.—On the State of Large Language Models and Applications to Manufacturing

Additive Manufacturing (AM), referred to as 3D printing, has emerged as a key pillar of Industry 4.0 enabling layer-by-layer fabrication of intricate geometries from CAD models. In parallel, Large Language Models (LLMs), deep learning models for natural language generation trained on vast text corpora, have demonstrated unprecedented capabilities in understanding and generating human-like text. The convergence of these trends opens new opportunities at the intersection of AM and AI/ML, where LLMs can assist engineers and researchers in design, manufacture planning, and knowledge discovery. Recent academic work has begun to explore LLM applications in AM and adjacent fields, such as material science, mechanical engineering, and design for additive manufacturing. This exploration ranges from intelligent process planning to domain-specific knowledge retrieval. This survey provides a comprehensive review of current developments, focusing on peer-reviewed literature contributions that apply, adapt, and advance LLMs in general and domain-specific domains. We analyze state-of-the-art (SOTA) techniques, such as fine-tuning foundational models for specific domains, retrieval-augmented generation (RAG) pipelines, knowledge graph integration, and delve into the architectures and evaluation methods employed. The goal of this survey is to inform researchers and practitioners of the current capabilities and limitations of LLMs in general and in domain-specific applications, and to outline how these models are being tailored to meet the requirements of these applications.

36 MATERIALS SCIENCE↗
NASA NTRS1990

Flight crew aiding for recovery from subsystem failures

Some of the conceptual issues associated with pilot aiding systems are discussed and an implementation of one component of such an aiding system is described. It is essential that the format and content of the information the aiding system presents to the crew be compatible with the crew's mental models of the task. It is proposed that in order to cooperate effectively, both the aiding system and the flight crew should have consistent information processing models, especially at the point of interface. A general information processing strategy, developed by Rasmussen, was selected to serve as the bridge between the human and aiding system's information processes. The development and implementation of a model-based situation assessment and response generation system for commercial transport aircraft are described. The current implementation is a prototype which concentrates on engine and control surface failure situations and consequent flight emergencies. The aiding system, termed Recovery Recommendation System (RECORS), uses a causal model of the relevant subset of the flight domain to simulate the effects of these failures and to generate appropriate responses, given the current aircraft state and the constraints of the current flight phase. Since detailed information about the aircraft state may not always be available, the model represents the domain at varying levels of abstraction and uses the less detailed abstraction levels to make inferences when exact information is not available. The structure of this model is described in detail.

Hudlicka, E.↗
NASA NTRSDate not supplied

Creating Benchmark Data for Artificial Intelligence and Machine Learning Space Biology Research

To identify an appropriate AI/ML approach for a specific problem, the best practice is to measure algorithm performance through the benchmarking process. A scientific benchmark consists of an AI-ready dataset and a reference implementation on a specific scientific question. The NASA Science Mission Directorate (SMD) has started the “Benchmark Initiative for AI/ML to create scientific benchmark datasets in three applications: 1) scientific benchmarking, which finds the best algorithm for a specific problem; 2) application benchmarking, which measures algorithm performance against a set of parameters; and 3) system benchmarking, which evaluates performance of hardware and software architecture. Currently, there are no standardized datasets available to benchmark AI/ML algorithms in the domain of space biology. In this work, we constructed two AI/ML-ready biological datasets from experiments in space-flown mice: cellular imaging and RNA-seq. First, radiation-exposed immune cells harbor DNA damage foci that can be fluorescently marked to visualize the amount of damage following exposure to ionizing radiation. However, such large datasets are difficult to analyze visually, due to imaging inconsistencies and human bias, and classical image processing approaches can fail on imaging artifacts. AI/ML are therefore exciting alternative, providing the speed of machines and the accuracy of humans. We have made this dataset available at https://registry.opendata.aws/bps_microscopy/. Second, high-throughput nucleic acid sequencing (DNA-seq, RNA-seq) has become widespread in biomedical research due to the growing availability and affordability of these assays. However, most sequencing datasets suffer from high dimensionality and low sample count. In this work, we used a generative adversarial network to synthesize a standardized, AI-ready, publicly available benchmark dataset for space biology RNA-seq data with sufficient space-flown and ground control mouse liver samples from NASA GeneLab. This dataset is available at https://registry.opendata.aws/bps_rnaseq/. These datasets are now fully open the Space Biology community to test their favorite AI/ML approaches.

James Casaletto↗
DOE OSTI2019

Active site plasticity and possible modes of chemical inhibition of the human DNA deaminase APOBEC3B

The single-stranded DNA cytosine deaminase APOBEC3B (A3B) functions in innate immunity against viruses, but it is also strongly implicated in eliciting mutations in cancer genomes. Because of the critical role of A3B in promoting virus and tumor evolution, small molecule inhibitors are desirable. However, there is no reported structure for any of the APOBEC3-family enzymes in complex with a small molecule bound in the active site, which hampers the development of small molecules targeting A3B. Here we report high-resolution structures of an active A3B catalytic domain chimera with loop 7 residues exchanged with those from the corresponding region of APOBEC3G (A3G). The structures reveal novel open conformations lacking the catalytically essential zinc ion, with the highly conserved active site residues extensively rearranged. These inactive conformations are stabilized by 2-pyrimidone or an iodide ion bound in the active site. Molecular dynamics simulations corroborate the remarkable plasticity of the engineered active site and identify key interactions that stabilize the native A3B active site. These data provide insights into A3B active site dynamics and suggest possible modes of its inhibition by small molecules, which would aid in rational design of selective A3B inhibitors for constraining virus and tumor evolution.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

The Story of 5d Metallocorroles: From Metal–Ligand Misfits to New Building Blocks for Cancer Phototherapeutics

Porphyrin chemistry is Shakespearean: over a century of study has not withered the field’s apparently infinite variety. Heme proteins continually astonish us with novel molecular mechanisms, while new porphyrin analogues bowl us over with unprecedented optical, electronic, and metal-binding properties. Within the latter domain, corroles occupy a special place, exhibiting a unique and rich coordination chemistry. The 5d metallocorroles are arguably the icing on that cake. New Zealand chemist Penny Brothers has used the word “misfit” to describe the interactions of boron, a small atom with a predilection for tetrahedral coordination, and porphyrins, classic square-planar ligands. Steve Jobs lionized misfits as those who see things differently and push humanity forward. Both perspectives have inspired us. The 5d metallocorroles are misfits in that they encapsulate a large 5d transition metal ion within the tight cavity of a contracted porphyrin ligand. Given the steric mismatch inherent in their structures, the syntheses of some 5d metallocorroles are understandably capricious, proceeding under highly specific conditions and affording poor yields. Three broad approaches may be distinguished. (a) In the metal–alkyl approach, a free-base corrole is exposed to an alkyllithium and the resulting lithio-corrole is treated with an early transition metal chloride; a variant of the method eschews alkyllithium and deploys a transition metal–alkyl instead, resulting in elimination of the alkyl group as an alkane and insertion of the metal into the corrole. This approach is useful for inserting transition metals from groups 4, 5, and, to some extent, 6, as well as lanthanides and actinides. (b) In our laboratory, we have often deployed a low-valent organometallic approach for the middle transition elements (groups 6, 7, 8, and 9). The reagents are low-valent metal–carbonyl or −olefin complexes, which lose one or more carbon ligands at high temperature, affording coordinatively unsaturated, sticky metal fragments that are trapped by the corrole nitrogens. (c) Finally, a metal acetate approach provides the method of choice for gold and platinum insertion (groups 10 and 11). This Account provides a first-hand perspective of the three approaches, focusing on the last two, which were largely developed in our laboratory. In general, the products were characterized with X-ray crystallography, electrochemistry, and a variety of spectroscopic methods. The physicochemical data, supplemented by relativistic DFT calculations, have provided fascinating insights into periodic trends and relativistic effects. An unexpected feature of many 5d metallocorroles, given their misfit character, is their remarkable stability under thermal, chemical, and photochemical stimulation. Many of them also exhibit long triplet lifetimes on the order of 100 μs and effectively sensitize singlet oxygen formation. Many exhibit phosphorescence in the near-infrared under ambient conditions. Furthermore, water-soluble ReO and Au corroles exhibit impressive photocytotoxicity against multiple cancer cell lines, promising potential applications as cancer phototherapeutics. We thus envision a bright future for the compounds as rugged building blocks for new generations of therapeutic and diagnostic (theranostic) agents.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Building MCP-native hierarchical AI scientist ecosystems: a perspective on scaling multi-agent scientific discovery

Large language models (LLMs) are evolving from chatbots with limited tool-using capabilities to agentic AI systems that can perform deep research, assist in proposing hypotheses, help design experiments, automate data analysis, and draft scientific reports. However, there are currently two bottlenecks limiting LLMs' real-world impact on the broader scientific research community beyond academic demonstrations: lack of interoperability (repetitive manual tool-integration is required across scenarios) and the need for scalable coordination (unstructured communication and memory become brittle as the number of agents grows). In this Perspective, we argue that the next phase of agentic scientific discovery requires the development of an ecosystem of protocol-native agents and tools organized through hierarchies inspired by human society, beyond the current paradigm of a single monolithic “AI scientist”. We use Model Context Protocol (MCP) as a concrete example of an emerging interoperability layer for scientific tool and context exchange, and we propose three complementary pathways to increase the scaling capabilities of an MCP-native scientific ecosystem by addressing the composability issues: (1) MCP servers for high-value scientific tools maintained by domain experts, (2) automated transformation of existing code repositories into MCP services, and (3) autonomous invention and evolution of new agents and workflows. Finally, we provide a practical roadmap for scaling AI-driven scientific discovery by expanding tool supply and coordination in MCP-native scientific ecosystems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

HEPTAPOD: Orchestrating High Energy Physics Workflows Towards Autonomous Agency

Many workflows in high-energy-physics (HEP) stand to benefit from recent advances in transformer-based large language models (LLMs). While early applications of LLMs focused on text generation and code completion, modern LLMs now support orchestrated agency: the coordinated execution of complex, multi-step tasks through tool use, structured context, and iterative reasoning. We introduce the HEP Toolkit for Agentic Planning, Orchestration, and Deployment (HEPTAPOD), an orchestration framework designed to bring this emerging paradigm to HEP pipelines. The framework enables LLMs to interface with domain-specific tools, construct and manage simulation workflows, and assist in common utility and data analysis tasks through schema-validated operations and run-card-driven configuration. To demonstrate these capabilities, we consider a representative Beyond the Standard Model (BSM) Monte Carlo validation pipeline that spans model generation, event simulation, and downstream analysis within a unified, reproducible workflow. HEPTAPOD provides a structured and auditable layer between human researchers, LLMs, and computational infrastructure, establishing a foundation for transparent, human-in-the-loop systems.

Menzo, Tony [Alabama U.; Fermilab] (ORCID:00000002↗
DOE OSTI2022

Machine learning with knowledge constraints for process optimization of open-air perovskite solar cell manufacturing

Perovskite photovoltaics (PV) have achieved rapid development in the past decade in terms of power conversion efficiency of small-area lab-scale devices; however, successful commercialization still requires further development of low-cost, scalable, and high-throughput manufacturing techniques. One of the critical challenges of developing a new fabrication technique is the high-dimensional parameter space for optimization, but machine learning (ML) can readily be used to accelerate perovskite PV scaling. Herein, we present an ML-guided framework of sequential learning for manufacturing process optimization. We apply our methodology to the Rapid Spray Plasma Processing (RSPP) technique for perovskite thin films in ambient conditions. With a limited experimental budget of screening 100 process conditions, we demonstrated an efficiency improvement to 18.5% as the best-in-our-lab device fabricated by RSPP, and we also experimentally found 10 unique process conditions to produce the top-performing devices of more than 17% efficiency, which is 5 times higher rate of success than the control experiments with pseudo-random Latin hypercube sampling. Our model is enabled by three innovations: (a) flexible knowledge transfer between experimental processes by incorporating data from prior experimental data as a probabilistic constraint; (b) incorporation of both subjective human observations and ML insights when selecting next experiments; (c) adaptive strategy of locating the region of interest using Bayesian optimization first, and then conducting local exploration for high-efficiency devices. Furthermore, in virtual benchmarking, our framework achieves faster improvements with limited experimental budgets than traditional design-of-experiments methods (e.g., one-variable-at-a-time sampling). This framework shows the capability of incorporating researchers’ domain knowledge into the ML-guided optimization loop; therefore, it has the potential to facilitate the wider adoption of ML in scaling to perovskite PV manufacturing.

14 SOLAR ENERGY↗
DOE OSTI2025

Increasing the Scale of the Mass Spectrometry Query Language Compendium with Explainable AI

A significant bottleneck in metabolomics data interpretation is the effective use of domain knowledge to assign structural information based on fragmentation patterns. The mass spectrometry query language (MassQL) aims to make this process accessible and applicable across multiple analysis platforms. While advanced computational methods are capable of predicting compound structures from fragmentation data, AI/ML approaches often rely on complex, opaque criteria that are difficult to interpret or modify. As a result, their predictive patterns cannot be readily translated into human-readable rules, such as those used in MassQL. Here, in this study, we introduce ChemEcho, a machine learning embedding method that converts tandem mass spectrometry data into sparse feature vectors containing peak and neutral mass subformulae to enhance explainable AI/ML-based methods. An advantage of this approach is that decision trees trained using these feature vectors can be directly translated to MassQL. Using a battery of decision trees trained using ChemEcho embeddings to predict molecular attributes, we generated over 1500 MassQL queries for 765 molecular features and evaluated their precision and recall. From these queries, the 50 highest-performing queries were integrated into the MassQL compendium. This set of generated MassQL queries included environmentally and biologically relevant classes such as PFAS and molecules containing phosphate or sulfate substructures. To illustrate the impact these queries would have on a typical metabolomics experiment, these MassQL queries were applied to a public metabolomics data set─resulting in a marked increase in the structural information derived from tandem mass spectra. Access and reuse of these queries is expected to enhance structural annotation in untargeted experiments, leading to more specific claims and advancing many applications in metabolomics.

Harwood, Thomas V. [USDOE Joint Genome Institute (↗
NASA NTRSDate not supplied

Communicating Metrics of Land Surface Temperature Variability Using Multi-sensor Machine Learning

Land surface temperature (LST) is a key climate observable used to detect changes in the Earth’s surface energy budget that influence carbon and water cycles. Land surface temperature exhibits strong diurnal variability, which geostationary satellites can observe at scale thanks to their temporal resolution. Due to anthropogenic climate and land use changes, the surface energy balance has been considerably modified and may be described by changes in diurnal temperature range and extremes. Using high performance computing and datasets from the NASA Earth Exchange, we exploit co-located, co-temporal observations from low-earth orbit (LEO) and geostationary (GEO) sensors to develop a deep learning-based method for LEO-to-GEO algorithm emulation. Our model is trained to predict MODIS Terra LST from GOES-16 thermal bands and achieves validation error <2K. Application of the model to unseen times of day (observed by MODIS Aqua) and a new GEO sensor (Himawari-8) observing an unseen spatial domain, demonstrate the generalization of the deep learning model across space, time and spectra. Further, time series clustering approaches are examined with the objective of identifying key indicators of change in diurnal cycling and extremes on a continental scale. Communicating LST variability observed by geostationary satellites can have impacts in multiple disciplines, from understanding of snow, vegetation and soil dynamics, to recognizing trends in heat events relevant to human health.

Kate Duffy↗
DOE OSTI2025

Data and Code for Understanding Generative AI Content with Embedding Models

This repository contains code for the experiments in the paper "Understanding Generative AI Content with Embedding Models". Constructing high-quality features is critical to any quantitative data analysis. While feature engineering was historically addressed by carefully hand-crafting data representations based on domain expertise, deep neural networks (DNNs) now offer a radically different approach. DNNs implicitly engineer features by transforming their input data into hidden feature vectors called embeddings. For embedding vectors produced by foundation models -- which are trained to be useful across many contexts -- we demonstrate that simple and well-studied dimensionality-reduction techniques such as Principal Component Analysis uncover inherent heterogeneity in input data concordant with human-understandable explanations. Of the many applications for this framework, we find empirical evidence that there is intrinsic separability between real samples and those generated by artificial intelligence (AI).

Vargas, Max [Pacific Northwest National Laboratory↗