Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data replication”

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 865 records · Page 48
All sourcesNASA NTRSDOE OSTI
DOE OSTI2014

SU-E-J-258: Inter- and Intra-Fraction Setup Stability and Couch Change Tolerance for Image Guided Radiation Therapy

Purpose: Immobilization devices combine rigid patient fixation as well as comfort and play a key role providing the stability required for accurate radiation delivery. In the setup step, couch re-positioning needed to align the patient is derived via registration of acquired versus reference image. For subsequent fractions, replicating the initial setup should yield identical alignment errors when compared to the reference. This is not always the case and further couch re-positioning can be needed. An important quality assurance measure is to set couch tolerances beyond which additional investigations are needed. The purpose of this work was to study the inter-fraction couch changes needed to re-align the patient and the intra-fraction stability of the alignment as a guide to establish the couch tolerances. Methods: Data from twelve patients treated on the Accuray CyberKnife (CK) system for fractionated intracranial radiotherapy and immobilized with Aquaplast RT, U-frame, F-Head-Support (Qfix, PA, USA) was used. Each fraction involved image acquisitions and registration with the reference to re-align the patient. The absolute couch position corresponding to the approved setup alignment was recorded per fraction. Intra-fraction set-up corrections were recorded throughout the treatment. Results: The average approved setup alignment was 0.03±0.28mm, 0.15±0.22mm, 0.06±0.31mm in the L/R, A/P, S/I directions respectively and 0.00±0.35degrees, 0.03±0.32degrees, 0.08±0.45degrees for roll, pitch and yaw respectively. The inter-fraction reproducibility of the couch position was 6.65mm, 10.55mm, and 4.77mm in the L/R, A/P and S/I directions respectively and 0.82degrees, 0.71degrees for roll and pitch respectively. Intra-fraction monitoring showed small average errors of 0.21±0.21mm, 0.00±0.08mm, 0.23±0.22mm in the L/R, A/P, S/I directions respectively and 0.03±0.12degrees, 0.04±0.25degrees, and 0.13±0.15degrees in the roll, pitch and yaw respectively. Conclusion: The inter-fraction reproducibility should serve as a guide to couch tolerances, specific to a site and immobilization. More patients need to be included to make general conclusions.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2026

New Report Highlights LAMP's Role as a Robust and Reliable System for Wave Energy Converter Testing

Harnessing the power of waves requires more than just putting devices in the ocean - it takes tools that can refine and accelerate innovation before the first deployment. That's where NLR's largeamplitude motion platform, or LAMP, comes in. LAMP is a six-degree-of-freedom motion system designed to replicate the response of a wave energy converter (WEC) device in realistic ocean conditions. By letting researchers evaluate WEC performance in a safe, controlled lab environment before moving offshore, LAMP helps accelerate innovation while reducing risk. An NLR report, titled Initial Characterization of the NLR LargeAmplitude Motion Platform (https://www.nrel.gov/docs/fy26osti/93733.pdf), provides a deeper understanding of LAMP's capabilities and also lays the groundwork for future advancements in WEC testing methodologies. Researchers studied LAMP's amplitude and frequency limits across different payloads and collected data from 16 test profiles. Testing demonstrated that platform accuracy remained consistent across all payloads evaluated; researchers expect similar accuracy up to LAMP's 10,000-kilogram capacity limit. The measured position limits for single degrees of freedom experienced by WECs, including surge, sway, heave, roll, pitch, and yaw, also exceeded the original manufacturer specifications, providing new insight into the true boundaries of the system.

16 TIDAL AND WAVE POWER↗
DOE OSTI2021

Assessment of a tomography-informed polyhedral discrete element modelling approach for complex-shaped granular woody biomass in stress consolidation

The design of handling equipment for granular biomass primarily requires experiments. Discrete element models (DEM) can provide designers with detailed insight into the behaviour of granular materials. However, granular biomass comprising complex-shaped particles is difficult to model with DEM. A tomography-informed DEM approach and an exhaustive assessment of the ability of the approach to predict the bulk behaviour of milled pines using experimental data is presented. Nano-CT scan was conducted to obtain 3D particle surface geometries as the basis for particle shape approximation by a polyhedral model and a sphero-polyhedral model. These models were applied in the simulation of a compressibility test. Our parametric study showed that particle Young's modulus and restitution coefficient are the two main properties influencing the simulated bulk behaviour of the DEM particles and that the level of approximating DEM particle shape for real particles is critical for accurately replicating the bulk behaviour of milled pines. The polyhedral model demonstrated better suitability than its sphero-polyhedral counterpart for modelling pine particles. The polyhedral model calibrated in a compressibility test was then applied in the simulations of a friction test without additional parameter tuning. Interlocking was dominant in the shear of bulk pine particles. Remarkably, the polyhedral model predicted the frictional behaviour of the pine particles when compared to the experimental results. The limitations of the model, as well as possible ways for enhancement, are discussed. This work provided novel insights into the suitability of complex-shaped DEM models for granular woody biomass.

Biomass flow↗
DOE OSTI2025

Scalability analysis of heavy-duty gas turbines using data-driven machine learning

With the increasing integration of variable renewable energy sources into power systems, the role of flexible power generation technologies like gas turbines (GT) in rapid grid balancing remains crucial. This sustained importance underscores the need for scaled and precise modeling of GT to ensure effective integration within evolving energy frameworks. While physics-driven GT models integrate thermodynamics, fluid dynamics, and combustion principles, they often rely on approximate mathematical representations to accommodate scaling that may not capture the actual complex dynamics for GTs and inertial effects associated to GTs with different ratings. In this study, a data-driven model is proposed using machine learning (ML) techniques to conduct GT scalability analysis and performance evaluation with high accuracy. The ML model, trained on data from various operating conditions and performance parameters, aims to uncover intricate relationships and patterns, resembling GT characteristics at different scales (ratings). The model is developed to capture complex system interaction and to adapt to changing operational scenarios at different capacities, providing valuable insights of power system dynamics. In this study, the real-time digital simulator platform was employed to generate training data for the ML model and assess its dynamic characteristics. The ultimate objective was to develop a detailed modeling framework based on governing equations and data-driven ML capable of predicting key performance indicators, in thermal systems such as GTs, including power output, speed, fuel consumption, and exhaust temperature under diverse operating conditions at different scales. The developed ML framework demonstrated high accuracy, with mean relative errors for GT power prediction, reference speed, exhaust temperature, and compressor pressure ratio (CPR) parameters consistently below 0.1% across typical load fluctuation scenarios. Maximum deviations were limited to approximately 0.5 K for exhaust temperature and 0.009 for CPR, underscoring the model’s ability to replicating dynamic GT behavior with high precision. The adaptability of the ML model enables its application across diverse operational conditions and its extension to other thermal systems. By leveraging advanced ML techniques, this study presents a robust and scalable modeling framework that enhances GT simulation precision, facilitating improved integration into evolving power systems.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2017

Enhanced Hanford Low-Activity Waste Glass Property Data Development: Phase 1

This report summarizes and analyzes data collected on a first test matrix of 36 low-activity waste (LAW) glass compositions intended to expand the LAW glass composition region over which glass property models are valid. The 36 LAW glass compositions were selected using statistical layered experimental design to explore a new glass composition region that overlaps and expands beyond the previously explored LAW glass composition region. The layered design of 36 LAW glasses consisted of 18 outer-layer glasses, 13 inner-layer glasses, a center glass (two replicates), and a glass previously tested at another laboratory (three replicates). The outer-layer glass compositions include extrema of the glass compositions that extend across and slightly beyond the compositions that are expected to be targeted for production. The analyses performed on these glasses include chemical composition (for target compositional verification), density, viscosity, electrical conductivity, crystal fraction, Canister Centerline Cooling with crystal identification, Product Consistency Test response, sulfur solubility, iron redox, and Vapor Hydration Test response. This report documents and discusses the results obtained from this testing.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2023

Measurement of Plutonium Ratios for CRM 137A Recertification

This report documents the analytical methodologies used by the Chemical and Isotopics Mass Spectrometry (CIMS) Group at the US Department of Energy’s Oak Ridge National Laboratory (ORNL) for plutonium isotopic ratio characterizations on purified ~1 mg units prepared by Lawrence Livermore National Laboratory (LLNL) from a mother solution created from an existing 250 mg unit of certified reference material (CRM) 137. The analytical processes described are designed to maintain traceability of the final certified attributes and to eliminate sources of systematic bias for the measurements described. The measurands for certification are the plutonium isotope ratios: 238 Pu/ 239 Pu, 240 Pu/ 239 Pu, 241 Pu/ 239 Pu, and 242 Pu/ 239 Pu. The preparation and analytical process included the following activities: Preparation of reference materials to be used as calibrants to establish traceability to the SI unit of mass for the isotopic analyses and quality controls. For the plutonium isotopic analyses, CRM 128 (a CRM comprising equal atoms of 239 Pu and 242 Pu) was used as the primary calibrant; Each CRM 137A unit was reconstituted using dilute high-purity HNO3. From each, seven replicates were distributed for the isotopic characterizations by thermal ionization mass spectrometry (TIMS) using the total evaporation (TE) method for isotope ratio measurements and MC-ICP-MS using a dynamic peak hooping method for secondary isotope ratio measurements; Due to time delay in the analyses from the last purification of the original CRM 137 mother solution by LLNL, an aliquot of the three provided plutonium units was purified using Eichrom TEVA separation methods to remove americium in-growth and uranium impurities; Minimally corrected data were reported from each analysis using reporting templates provided by the New Brunswick Laboratory Program Office (NBL PO) for the specific methods used for isotopic measurements. In addition, a copy of the instrument exports, Excel templates for isotopic data corrections, and the uncertainty budgets completed with the GUM Workbench software package were submitted to NBL PO.

07 ISOTOPE AND RADIATION SOURCES↗
DOE OSTI2018

Pol I DNA polymerases stimulate DNA end-joining by Escherichia coli DNA ligase

Highlights: • Pol I DNA polymerases stimulate E. coli DNA ligase activity. • End-joining activity is enhanced by up to 600%, nick closure is slightly enhanced. • T4 ligase is not similarly effected. Klenow and Klentaq are the large fragment domains of the Pol I DNA polymerases from Escherichia coli and Thermus aquaticus, respectively. Herein, we show that both polymerases can significantly stimulate complementary intermolecular end-joining ligations by E.coli DNA ligase when the polymerases are present at concentrations lower than that of the DNA substrates. In contrast, high polymerase concentrations relative to the DNA substrates inhibit the intermolecular ligation activity of DNA ligase. Neither polymerase was able to stimulate the DNA ligase from T4 bacteriophage. Additionally, nick-closure by E. coli DNA ligase (but not T4 ligase) is slightly stimulated by both polymerases, but only at about 10% of the magnitude seen for end-joining enhancement. The data represent one of the first observations of direct polymerase-ligase interactions in prokaryotes, and suggest that the polymerases stabilize the associated DNA ends during intermolecular ligation, and that such a complex can be taken advantage of by some, but not all, DNA ligases.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

A phage nucleus-associated RNA-binding protein is required for jumbo phage infection

Abstract Large-genome bacteriophages (jumbo phages) of the proposed family Chimalliviridae assemble a nucleus-like compartment bounded by a protein shell that protects the replicating phage genome from host-encoded restriction enzymes and DNA-targeting CRISPR-Cas nucleases. While the nuclear shell provides broad protection against host nucleases, it necessitates transport of mRNA out of the nucleus-like compartment for translation by host ribosomes, and transport of specific proteins into the nucleus-like compartment to support DNA replication and mRNA transcription. Here, we identify a conserved phage nuclear shell-associated protein that we term Chimallin C (ChmC), which adopts a nucleic acid-binding fold, binds RNA with high affinity in vitro, and binds phage mRNAs in infected cells. ChmC also forms phase-separated condensates with RNA in vitro. Targeted knockdown of ChmC using mRNA-targeting dCas13d results in accumulation of phage-encoded mRNAs in the phage nucleus, reduces phage protein production, and compromises virion assembly. Taken together, our data show that the conserved ChmC protein plays crucial roles in the viral life cycle, potentially by facilitating phage mRNA translocation through the nuclear shell to promote protein production and virion development.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

ECP ST Capability Assessment Report (CAR) for VTK-m (FY20)

The ECP/VTK-m project is providing the core capabilities to perform scientific visualization on Exascale architectures. The ECP/VTK-m project fills the critical feature gap of performing visualization and analysis on processors like graphics-based processors. The results of this project will be delivered in tools like ParaView, Vislt, and Ascent as well as in stand-alone form. Moreover, these projects are depending on this ECP effort to be able to make effective use of ECP architectures. One of the biggest recent changes in high-performance computing is the increasing use of accelerators. Accelerators contain processing cores that independently are inferior to a core in a typical CPU, but these cores are replicated and grouped such that their aggregate execution provides a very high computation rate at a much lower power. Current and future CPU processors also require much more explicit parallelism. Each successive version of the hardware packs more cores into each processor, and technologies like hyper threading and vector operations require even more parallel processing to leverage each core's full potential. VTK-m is a toolkit of scientific visualization algorithms for emerging processor architectures. VTK-m supports the fine-grained concurrency for data analysis and visualization algorithms required to drive extreme scale computing by providing abstract models for data and execution that can be applied to a variety of algorithms across many different processor architectures. The ECP/VTK-m project is building up the VTK-m codebase with the necessary visualization algorithm implementations that run across the varied hardware platforms to be leveraged at the Exascale. We will be working with other ECP projects, such as ALPINE, to integrate the new VTK-m code into production software to enable visualization on our HPC systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Materials Characterization, Prediction, and Control Project: Summary Report on Material Characterization, Part 1

The Pacific Northwest National Laboratory (PNNL) undertook the Materials Characterization, Prediction, and Control (MCPC) Laboratory Directed Research and Development Project to advance understanding of nuclear material processing and enable multifold acceleration in the development and qualification of new material systems in national security and advanced energy applications (Smith 2021). The MCPC Project executed research across three scientific vertices—material characterization, predictive modeling, and data analytics—with extensive support by a data curation and management team. The central technical objective in the MCPC Project was to improve the prediction and characterization of the process-structure-property relationships within the microstructurally refined region of stainless-steel samples prepared utilizing friction stir processing (FSP). Application of the FSP technique is well established at PNNL within the Solid Phase Processing capability through many years of investment across a range of materials and applications (PNNL 2024). Three distinct rounds of FSP experiments were performed by the experimental team, producing replicate samples utilizing across different nominal processing conditions (Condition IDs) listed in Table 1. The starting material on which FSP was applied was commercially available unprocessed stainless-steel type 316L material. Chosen processing conditions were very diverse, and some were intentionally chosen to produce defects. Several samples experienced tool breakage during experimentation, so a full set of three replicates was not produced for every nominal processing condition.

36 MATERIALS SCIENCE↗
DOE OSTI2014

Porcine circovirus-2 capsid protein induces cell death in PK15 cells

Studies have shown that Porcine circovirus (PCV)-2 induces apoptosis in PK15 cells. Here we report that cell death is induced in PCV2b-infected PK15 cells that express Capsid (Cap) protein and this effect is enhanced in interferon gamma (IFN-γ)-treated cells. We further show that transient PCV2a and 2b-Cap protein expression induces cell death in PK15 cells at rate similar to PCV2 infection, regardless of Cap protein localization. These data suggest that Cap protein may have the capacity to trigger different signaling pathways involved in cell death. Although further investigation is needed to gain deeper insights into the nature of the pathways involved in Cap-induced cell death, this study provides evidence that PCV2-induced cell death in kidney epithelial PK15 cells can be mapped to the Cap protein and establishes the need for future research regarding the role of Cap-induced cell death in PCV2 pathogenesis. - Highlights: • IFN-γ enhances PCV2 replication that leads to cell death in PK15 cells. • IFN-γ enhances nuclear localization of the PCV2 Capsid protein. • Transient PCV2a and 2b-Capsid protein expression induces cell death. • Cell death is not dictated by specific Capsid protein sub-localization.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Liquid–Liquid Equilibrium Prediction in Fast Pyrolysis Bio-Oil Systems: A Framework for Incorporating Bio-Oil Complexity

The study of mixtures of bio-oil, water and organic solvents in different proportions can serve as a cost-effective analysis of its content due to the formation of immiscible phases. This manuscript attempts to replicate experimentally determined partition coefficients (K OW ) of relevant species present in fast pyrolysis bio-oil (FPBO). A commercial flowsheeting simulator with surrogate bio-oil model representation is used. Concurrently, pyrolytic lignins in FPBO (‘pyrolignin’) do not have an agreed-upon structural representation, and the literature is ripe with wide variations of said representations. Thus, during the description of FPBO, this pyrolignin fraction was modeled using 20 possible structures (phenolic dimers to tetramers), with the goal of determining the structures for which the experimental data are best described. Two cases were considered: Case 1 normalized the reported experimental mass balance, while Case 2 included the unreported fraction in the mass balance to the total pyroligin. Please, add here a comment on the prediction of the Water oil equilibrium. The best KOW predictions for levoglucosan (LVG) were obtained when the system was modeled with no pyrolignin, presenting an MRE under 10% for both systems WO and BO. Among the possible structures, D2 (dimer), F1 (trimer) and I1, and I3 (tetramers) presented MRE ≤ 13% for both cases.

09 BIOMASS FUELS↗
DOE OSTI2022

Coincident Capture through Post-processing PTRAC [Slides]

This presentation discusses the new PTRAC capabilities and workflows. The PTRAC capability in MCNP6.3 has seen a massive overhaul since MCNP6.2. The new HDF5 file format allows for both MPI- and thread-based parallelism. MCNPTools has been updated to handle the new HDF5 PTRAC format and is now open sourced on GitHub. Built-in capabilities, such as the pulse-height tally coincident capture special treatment, can largely be replicated through separate postprocessing scripts that leverage both PTRAC and MCNPTools. This allows for greater flexibility in user-specified and controlled detector response functionality, ultimately using the MCNP code for what it is best at (i.e., particle transport).

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2019

PRIMUS: Clustering of Star-forming and Quiescent Central Galaxies at 0.2 < z < 0.9

Previous work has demonstrated that at a given stellar mass, quiescent galaxies are more strongly clustered than star-forming galaxies. The contribution to this signal from central, as opposed to satellite, galaxies is not known, which has strong implications for galaxy evolution models. To investigate the contribution from central galaxies, we present in this work measurements of the clustering of isolated primary (IP) galaxies, used as a proxy for central galaxies, at 0.2 < z < 0.9 with data from the PRIMUS galaxy redshift survey. Using a sample of spectroscopic redshifts for ~ 60, 000 galaxies with M * ≳ 10 9 M ⊙ covering 5 deg 2 on the sky, we define IP galaxies using isolation cuts in spatial proximity and stellar mass of nearby galaxies. We find that at fixed stellar mass, quiescent IP galaxies are more strongly clustered than star-forming IP galaxies at z ~ 0.35 (10σ). Using mock galaxy catalogs based on recent halo occupation models of Behroozi et al. (2019) and designed to replicate the parameters of the PRIMUS survey dataset, we find that these clustering differences are due in part to quiescent central galaxies being more strongly clustered than star-forming central galaxies. This is consistent with either distinct stellar-to-halo mass relations for quiescent and star-forming central galaxies, and/or central galaxy assembly bias. We additionally use mock catalogs to assess the dependence of both incompleteness and satellite galaxy contamination in the IP galaxy samples on redshift, galaxy type, and stellar mass, and demonstrate how isolation criteria yield biased subsamples of central galaxies via environmental incompleteness, or the preferential exclusion of central galaxies in overdense environments.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Effects of applied stress and grain size on creep-rupture lifetime prediction for Haynes 282 alloy

A multiple-heat set of tensile and creep-rupture data generated from a single laboratory was used to establish and evaluate predictive equations for the average creep-limited lifetime of an aged wrought Haynes 282 alloy using two different time- and temperature-compensated parametric models (those of Larson and Miller and Wilshire et al.). The range of temperatures and stresses used to produce the creep-rupture data used in the analyses revealed that two distinct regimes, above and below the proportional limit, were necessary for more accurate predictions of creep lifetimes over the range of conditions studied. Furthermore, a 2x difference in grain size amongst the three heats revealed a breakdown in the relationship between ultimate tensile strength and creep-rupture time, thus impacting the predictive ability of the Wilshire model. Finally, lifetime predictions based on the formulations from the respective models were compared to other available Haynes 282 datasets and found to reasonably replicate the experimental trends established in other studies.

36 MATERIALS SCIENCE↗
DOE OSTI2013

Hydrogen uptake in Zircaloy-2 reactor fuel claddings studied with elastic recoil detection

The recent trend towards a high burn-up discharge spent nuclear fuel necessitates a thorough understanding of hydrogen uptake in Zr-based cladding materials that encapsulate spent nuclear fuel. Although it is challenging to experimentally replicate exact conditions in a nuclear reactor that lead to hydrogen uptake in claddings, in this study we have attempted to understand the kinetics of hydrogen uptake by first electrolytically charging Zircaloy-2 (Zr-2) cladding material for various durations (100 to 2,600 s), and subsequently examining hydrogen ingress with elastic recoil detection (ERD) and transmission electron microscopy (TEM). To understand the influence of irradiation damage defects on hydrogen uptake, an analogous study was performed on ion - irradiated (0.1, 1 and 25 dpa) Zr-2. Analysis of ERD data from the un-irradiated Zr-2 suggests that the growth of the hydride layer is diffusion controlled, and preliminary TEM results support this assertion. In un-irradiated Zr-2, the diffusivity of hydrogen in the hydride phase was found to be approximately 1.1 Multiplication-Sign 10{sup -11} cm{sup 2}/s, while the diffusivity in the hydride phase for lightly irradiated (0.1 and 1 dpa) Zr-2 is an order of magnitude lower. Irradiation to 25 dpa results in a hydrogen diffusivity that is comparable to the un-irradiated Zr-2. These results are compared with existing literature on hydrogen transport in Zr - based materials.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2018

Emergence and selection of isoniazid and rifampin resistance in tuberculosis granulomas

Drug resistant tuberculosis is increasing world-wide. Resistance against isoniazid (INH), rifampicin (RIF), or both (multi-drug resistant TB, MDR-TB) is of particular concern, since INH and RIF form part of the standard regimen for TB disease. While it is known that suboptimal treatment can lead to resistance, it remains unclear how host immune responses and antibiotic dynamics within granulomas (sites of infection) affect emergence and selection of drug-resistant bacteria. We take a systems pharmacology approach to explore resistance dynamics within granulomas. We integrate spatio-temporal host immunity, INH and RIF dynamics, and bacterial dynamics (including fitness costs and compensatory mutations) in a computational framework. We simulate resistance emergence in the absence of treatment, as well as resistance selection during INH and/or RIF treatment. There are four main findings. First, in the absence of treatment, the percentage of granulomas containing resistant bacteria mirrors the non-monotonic bacterial dynamics within granulomas. Second, drug-resistant bacteria are less frequently found in non-replicating states in caseum, compared to drug-sensitive bacteria. Third, due to a steeper dose response curve and faster plasma clearance of INH compared to RIF, INH-resistant bacteria have a stronger influence on treatment outcomes than RIF-resistant bacteria. Finally, under combination therapy with INH and RIF, few MDR bacteria are able to significantly affect treatment outcomes. Overall, our approach allows drug-specific prediction of drug resistance emergence and selection in the complex granuloma context. Since our predictions are based on pre-clinical data, our approach can be implemented relatively early in the treatment development process, thereby enabling pro-active rather than reactive responses to emerging drug resistance for new drugs. Furthermore, this quantitative and drug-specific approach can help identify drug-specific properties that influence resistance and use this information to design treatment regimens that minimize resistance selection and expand the useful life-span of new antibiotics.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2019

Defending Against Adversarial Examples

Adversarial machine learning is an active field of research that seeks to investigate the security of machine learning methods against cyber-attacks. An important branch of this field is adversarial examples, which seek to trick machine learning models into misclassifying inputs by maliciously tampering with input data. As a result of the pervasiveness of machine learning models in diverse areas such as computer vision, health care, and national security, this vulnerability is a rapidly growing threat. With the increasing use of AI solutions, threats against AI must be considered before deploying systems in a contested space. Adversarial machine learning is a problem strongly tied to software security, and just like other more common software vulnerabilities, it exploits a weakness in software, like components of machine learning models. During this project, we attempted to survey and replicate several adversarial machine learning techniques with the goal of developing capabilities for Sandia to advise and defend against these threats. To accomplish this, we scanned state of the art research for robust defenses against adversarial examples and applied them to a machine learning problem.

97 MATHEMATICS AND COMPUTING↗