Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Real numbers”

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 721 records · Page 40
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Assessment of human nuclear and mitochondrial DNA qPCR assays for quantification accuracy utilizing NIST SRM 2372a

In forensic DNA casework, a highly accurate real-time quantitative polymerase chain reaction (qPCR) assay is recommended per the Scientific Working Group on DNA Analysis Methods (SWGDAM) (SWGDAM Validation Guidelines for DNA Analysis Methods [1]) to determine whether a DNA sample is of sufficient quantity and robust quality to move forward with downstream short tandem repeats (STR) or sequencing analyses. Most of these assays rely on a standard curve, referred to herein and traditionally as absolute qPCR, in which an unknown is compared, relative to that curve. However, one fundamental issue with absolute qPCR is the quantifiable concentration of commercial assay standards can vary depending on (1) origin, i.e., whether from a cell line or a human subject, (2) supplier, (3) lot number, (4) shipping method, etc. In 2018, the National Institute for Standards and Technology (NIST) released a human DNA standard reference material for evaluating qPCR quantification standards, Standard Reference Material (SRM) 2372a, Romsos et al. (2018) [2] which contains three well-characterized human genomic DNA samples: Component A) a single male1 donor, Component B) a single female 1 donor, and Component C) a 1:3 male 2 :female 2 donor, each with certification data for nDNA and informational mitochondrial DNA(mtDNA)/nuclear DNA (nDNA) ratio data. The SRM 2372a was used to assess four qPCR assays: (1) Quantifiler Trio (Thermo Fisher Scientific, Waltham, MA) for nDNA quantification, (2) NovaQUANT (EMD Millipore Corporation, San Diego, CA) for nDNA and mtDNA quantification, (3) a custom duplex mtDNA assay, and (4) a custom triplex mtDNA assay. Additionally, extracts from eighteen (18) skeletal remains were tested with the latter three assays for concordance of DNA concentration and with assays (2) and (3), for the degradation state. Our assessment revealed that an accurate, efficient, and reproducible qPCR assay is dependent on (1) the quality and reliability of the DNA standard, (2) the qPCR chemistry, and (3) the specific primers, and probes (if applicable), used in an assay. Finally, our findings indicate qPCR assays may not always quantify as expected and that performance of each lot should be verified using a well-characterized DNA standard such as the NIST SRM 2372a and adjusted if warranted.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

A deep learning-based direct forecasting of CO 2 plume migration

Accurate and timely forecasts of CO 2 plume evolution in geological reservoirs are crucial for CO 2 migration detection, leakage risk assessment, and operation decision support. Conventional forecasting usually adopts a two-step strategy, first calibrating reservoir model parameters against observations using iterative inverse modeling (or history matching) and then applying the calibrated model for predictions. This method impedes real-time forecasts due to the heavy computational demand in inverse modeling and may suffer from poor prediction accuracy because of the limited observation data. In this work, we propose a deep learning-based latent space mapping framework to forecast CO 2 plume migration directly by avoiding the inverse modeling. We first use the convolutional autoencoder to map the high-dimensional complex plume extents onto low-dimensional latent space. Next, we use neural networks to learn the relationship between the observation variables and the prediction latent variables. And then for given observation data, we infer the prediction values directly. This one-step direct forecasting is computationally efficient which requires a few number of parallelizable reservoir simulations and it can provide accurate predictions with limited observations by learning the observation-prediction relationship in the reduced dimension. Therefore, our proposed method enables an in-time forecast of dynamic CO 2 plume distributions. In this work, we demonstrate the effectiveness and accuracy of our method in predicting the CO 2 plume migration using four metrics such as plume area, centroid movement distance, and plume spreading in the primary and secondary directions. And the spatio-temporal evolution patterns of plume migration under diverse geological complexities are also accurately quantified.

15 GEOTHERMAL ENERGY↗
DOE OSTI2020

Autonomous Tuning and Charge-State Detection of Gate-Defined Quantum Dots

Defining quantum dots in semiconductor-based heterostructures is an essential step in initializing solid-state qubits. With growing device complexity and increasing number of functional devices required for measurements, a manual approach to finding suitable gate voltages to confine electrons electrostatically is impractical. Here, we implement a two-stage device characterization and dot-tuning process, which first determines whether devices are functional and then attempts to tune the functional devices to the single or double quantum-dot regime. We show that automating well-established manual-tuning procedures and replacing the experimenter’s decisions by supervised machine learning is sufficient to tune double quantum dots in multiple devices without premeasured input or manual intervention. The quality of measurement results and charge states are assessed by four binary classifiers trained with experimental data, reflecting real device behavior. We compare and optimize eight models and different data preprocessing techniques for each of the classifiers to achieve reliable autonomous tuning, an essential step towards scalable quantum systems in quantum-dot-based qubit architectures.

Darulova, J↗
DOE OSTI2024

Enhancing quantum utility: Simulating large-scale quantum spin chains on superconducting quantum computers

We present the quantum simulation of the frustrated quantum spin- 1 2 antiferromagnetic Heisenberg spin chain with competing nearest-neighbor ( J 1 ) and next-nearest-neighbor ( J 2 ) exchange interactions in the real superconducting quantum computer with qubits ranging up to 100. In particular, we implement the Hamiltonian with the next-nearest neighbor exchange interaction in conjunction with the nearest-neighbor interaction on IBM's superconducting quantum computer and carry out the time evolution of the spin chain by employing the first-order Trotterization. Furthermore, our implementation of the second-order Trotterization for the isotropic Heisenberg spin chain, involving only nearest-neighbor exchange interaction, enables precise measurement of the expectation values of staggered magnetization observable across a range of up to 100 qubits. Notably, in both cases, our approach results in a constant circuit depth in each Trotter step, independent of the number of qubits. Our demonstration of the accurate measurement of expectation values for the large-scale quantum system using superconducting quantum computers designates the quantum utility of these devices for investigating various properties of many-body quantum systems. This will be a stepping stone to achieving the quantum advantage over classical ones in simulating quantum systems before the fault tolerance quantum era. Published by the American Physical Society 2024

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Dissolved oxygen sensor in an automated hyporheic sampling system reveals biogeochemical dynamics

Many river corridor systems frequently experience rapid variations in river stage height, hydraulic head gradients, and residence times. The integrated hydrology and biogeochemistry of such systems is challenging to study, particularly in their associated hyporheic zones. Here we present an automated system to facilitate 4-dimensional study of dynamic hyporheic zones. It is based on combining real-time in-situ and ex-situ measurements from sensor/sampling locations distributed in 3-dimensions. A novel dissolved oxygen (DO) sensor was integrated into the system during a small scale study. We measured several biogeochemical and hydrologic parameters at three subsurface depths in the riverbed of the Columbia River in Washington State, USA, a dynamic hydropeaked river corridor system. During the study, episodes of significant DO variations (~+/- 4 mg/l) were observed, with minor variation in other parameters (e.g., <~+/-0.15 mg/l NO 3 ). DO concentrations were related to hydraulic head gradients, showing both hysteretic and non-hysteretic relationships with abrupt (hours) transitions between the two types of relationships. The observed relationships provide a number of hypotheses related to the integrated hydrology and biogeochemistry of dynamic hyporheic zones. We suggest that preliminary high-frequency monitoring is advantageous in guiding the design of long term monitoring campaigns. The study also demonstrated the importance of measuring multiple parameters in parallel, where the DO sensor provided the key signal for identifying/detecting transient phenomena.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

A Clustering-Based Scenario Generation Framework for Power Market Simulation with Wind Integration

A critical step in stochastic optimization models of power system analysis is to select a set of appropriate scenarios and significant numbers of scenario generation methods exist in the literature. This paper develops a clustering based scenario generation method, which aims to improve the performance of existing scenario generation techniques by grouping a set of correlated wind sites into clusters according to their cross-correlations. Copula based models are utilized to model spatiotemporal correlations and the Gibbs sampling is then used to generate scenarios for day-ahead markets. Our results show that the generated scenarios based on clustered wind sites outperform existing approaches in terms of reliability and sharpness and can reduce the total computational time for scenario generation and reduction significantly. The clustering-based framework can therefore provide a better support for real-world market simulations with high wind penetration.

data visualization↗
DOE OSTI2025

Application of Raman Spectroscopy to Determine Uranium Content in ADUN Solution

The work presented in this report is part of the ongoing efforts to address the nuclear material control and accounting needs for advanced reactor fuel fabrication facilities. This work was supported by the Materials Protection, Accounting, and Control Technologies (MPACT) program under the US Department of Energy Office of Nuclear Energy‘s Nuclear Fuel Cycle and Supply Chain program. The activities and engagements under the MPACT program are designed to support a robust US civilian nuclear energy enterprise. In the work described in this report, we supported MPACT objectives by developing measurement techniques that could be used for material accounting and process monitoring and by working with industry partners to identify existing gaps and areas for improvement. Oak Ridge National Laboratory has been working with commercial tristructural isotropic (TRISO) fuel fabricators such as Standard Nuclear to develop technology for rapid and cost-effective uranium content assessment. This work has focused on demonstrating advanced measurement techniques (e.g., Raman spectroscopy) that can be used for rapid, reliable, and cost-effective routine measurements of uranium content in feed solutions and liquid waste streams as well as for monitoring in-line process measurements and product streams. Specifically, this report explores techniques for accurately determining uranium content in acid-deficient uranyl nitrate (ADUN) solutions and detecting low uranium concentrations in ammonia solutions. Developing such measurement techniques will benefit TRISO fuel fabrication facilities, facilities involved in other parts of the fuel cycle that require online monitoring of aqueous solutions, and potentially molten salt fuel reactors. This work supports developing Raman spectroscopy procedures to determine uranium concentrations in ADUN solutions, which are used as feedstock in the sol–gel process for creating TRISO fuel. Some additional benefits of using Raman spectroscopy for uranium quantification in fabrication facilities include enabling online monitoring of the chemical process, which would provide near real-time feedback; eliminating the need for sample transfers, preparation, or dilution; providing nondestructive measurements; and user friendliness. In this fiscal year, FY25, we determined the identity of the unknown Raman band at approximately 853 cm−1 that was discovered in ADUN Raman spectra in FY24, created calibration curves and determined uranium concentrations of two ADUN solutions, and compared the Raman results to results obtained from inductively coupled plasma mass spectrometry and Davies–Gray titration. Furthermore, we have identified focus areas for experimentation in future fiscal years. A key result is that the accuracy of using Raman can provide accuracy comparable to destructive analysist techniques, With a well-developed calibration curve, using standards and a large number of samples (more than five samples), uncertainty on the order of 1%–3% is achievable. Given that the uncertainties achieved by Raman spectroscopy were on the order of the uncertainties achieved using ICP-MS, we conclude that with a well-developed procedure Raman spectroscopy can be used to determine uranium concentrations in ADUN solutions for NMC&A applications. The benefits of such an approach are that the time and effort will be less than that of comparable destructive analysis techniques, with approximately the same level of technical expertise. This will be attractive to operators of fuel fabrication facilities as it will lower costs and improve efficiencies.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

Handling and Properties of Methanol as a Marine Fuel

Given the increasing concern around greenhouse gas emissions and the decline in the availability of fossil fuels, there is increasing global demand to develop alternate fuels for maritime transportation that are sustainable and which have lower greenhouse gas emissions. Methanol is one such alternative fuel that has garnered considerable attention given its potential to be produced by more sustainable processes and its more favorable greenhouse gas emission profile in comparison with current fossil fuels. Understanding the physical and chemical properties of methanol under a range of conditions is essential for its development as a marine fuel. In this study, we seek to define physical and chemical properties of different methanol samples to simulate real-world storage conditions as these data are lacking in the literature. Several methanol samples were evaluated: nearly pure methanol; International Organization for Standardization (ISO) marine methanol (MM) grades A, B, and C; and methanol plus higher alcohols. We first evaluated all methanol samples for impurities, acetic acid content, density, and distillation range. We then characterized the effects of water absorption and found that methanol can easily absorb unacceptable water content from humid air within hours, necessitating storage conditions that prevent this process. In eight-week aging experiments at 20 °C and 40 °C in ambient air, we did not observe significant oxidation for any of the methanol samples; however, we did observe increases in acid number. We assessed the impact of contamination of methanol with water, marine gas oil (MGO), and an MGO–biodiesel mixture on density, viscosity, distillation range, and lubricity. Finally, we show that MGO contamination of methanol results in a slight increase in sooting tendency. In aggregate, our results provide an in-depth analysis of physical and chemical properties of methanol as well as the impacts of storage conditions and impurities on the properties of fuel methanol.

09 BIOMASS FUELS↗
DOE OSTI2023

Proxy-based Bayesian inversion of strain tensor data measured during well tests

Recent instrument developments have made it possible to measure the strain tensor caused by injecting or pumping fluid from aquifers or reservoirs, but the full value of these data is limited because the long runtimes of poroelastic forward models makes it impractical to use many inversion schemes. This limits the interpretation of strain data for managing the recovery of resources or storage of wastes in the subsurface. This paper describes a method of inverting deformation data using a poroelastic numerical simulator so the results can be used to manage reservoirs or aquifers. We developed a workflow designed to reduce the number of simulations sufficiently to make it feasible to use DREAMzs, an advanced Bayesian inversion method that translates the uncertainties from different sources into unbiased posterior parameter distributions and uncertainty envelopes around the field data. Using a KNN proxy model for the poroelastic simulator is key to reducing the overall computations, and the workflow includes a strategy for ensuring the proxy model results converge on the results from the simulator. The workflow is tested using an idealized example that verifies the ability to correctly identify parameters and characterize noise used to perturb the data. Field data from an injection test at an oil reservoir near Tulsa, Oklahoma, are also used to evaluate the efficacy of the workflow with a real dataset. The workflow identified 265 history matching solutions out of 1240 total simulation runs (21% acceptance ratio), where the results were used to characterize posterior parameter distribution and evaluate the prediction uncertainty. Furthermore, this workflow is significant because it enables strain tensor, or other geomechanical measurements to be interpreted to guide decision-making during energy and environmental processes in the subsurface.

42 ENGINEERING↗
DOE OSTI2025

Boosting Noise2Inverse via enhanced model selection for denoising computed tomography data

Synchrotron-based x-ray tomographic imaging enables the examination of the internal structure of materials at high spatial and temporal resolution. Experimental constraints can impose dose and time limits on the measurements, introducing a higher level of noise and artifacts in the reconstructed images. Deep learning has emerged as a powerful tool to remove noise from reconstructed images. Recently, the Noise2Inverse method was designed specifically for denoising reconstructed images without requiring paired noisy and clean images. This method creates multiple statistically independent reconstructions used to pair the data in which training involves transforming one reconstruction into the other, and vice versa. Originally designed to be used after a fixed number of epochs, we see in practice that this approach may not produce the optimal model and may unnecessarily waste computational resources. Therefore, we propose an alternative method of identifying the best model during training that aligns with the Noise2Inverse method. During validation, we compare the model output of the multiple reconstructions among each other. We hypothesize that the best model is the one that produces images with the highest similarity, implying a convergence in the predicted material properties and absorption values. To compare model outputs, we consider the absolute error, square error, structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and cosine similarity. We evaluate our method on two simulated tomography datasets and two, real-world, low-contrast, high-energy x-ray tomography datasets. We show our approach is more effective at determining the best model, up to an increase of 12.50% and 12.53% in SSIM and PSNR, respectively, while only requiring a fifth of the training time compared to the original approach.

CT↗
DOE OSTI2022

Machine Learning for Automated Metadata Assignment in Buildings: Cooperative Research and Development (Final Report, CRADA Number CRD-18-00767)

RealTerm Energy and NREL have identified a shared vision to evaluate opportunities to facilitate the organization and assignment of metadata to building control system (BCS) data via industry-informed machine learning (ML). Manual metadata assignment is labor intensive and costly, slowing down any Energy Management and Information System (EMIS) deployment in the building space. This project aims to develop methodologies to accurately assign this metadata and significantly decrease the level of effort associated with deploying EMIS. The objective of this project is to identify/design methodologies to assign metadata to HVAC control points automatically. The identified methodologies will be programmed in analytics algorithms so they can ingest a list of points and produce a detailed tagging following the Haystack classification nomenclature. To validate the efficacy of each methodology, tagging results will be compared utilizing a list of points extracted from RealTerm's building database-as well as data extracted from the NREL campus via the Intelligent Campus program-enabling testing against large datasets with real world challenges. The developed methodologies may leverage building manager/operator input on a limited basis to add context to the classifying algorithms. The partnership aims to advance global efforts in areas related to the DOE missions through improving operational performance of commercial buildings. It is well documented that buildings fall out of commission after they are occupied, wasting significant energy and incurring associated costs simply due to poor operational performance. Emerging EMIS technologies that perform continuous commissioning help to address this issue, yet integration of these systems can be labor intensive both for the technology vendor and the building owner/operator. This project will enable more efficient and cost-effective analytics for buildings, enabling improvement in building operations at lower cost points.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2026

FOCAL Campaign I: Advanced Wind Turbine Control Strategies

Campaign I of the Floating Offshore-wind Controls Advanced Laboratory Experimental Program (FOCAL) aims to generate a dataset enabling the validation of aerodynamic performance of a scaled turbine mounted on a rigid tower in a fixed condition. The turbine considered in the FOCAL testing campaigns is the IEA-Wind 15MW Reference Wind Turbine. This scaled model is capable of simulating advanced blade-pitch control strategies in a high-quality wind field. The turbine is fully instrumented to record a variety of parameters in real time such as structural loads and dynamics. The test data considered was generated at the University of Maine's Harold Alfond Wind and Wave (W2) testing facility. The Load Cases (LC) considered in this testing campaign are as follows: LC 1.X - Constant wind and blade-pitch with varying rotor speeds LC 2.X - Constant wind and rotor speed with varying blade-pitch LC 3.X - Varying wind with active closed-loop control Detailed properties on the modeled system are found in the following reference: Lenfest E., Floating Offshore-wind Controls Advanced Laboratory (FOCAL) Experimental Program - Campaign I: 1:70 Model-scale Testing of the IEA-Wind 15MW Reference Turbine. UMaine ASCC Report Number 23-40-1183. Details on the results of the verification and validation are found in the following reference: Mendoza, Nicole et al., "Verification and Validation of Model-Scale Turbine Performance and Control for the IEA Wind 15 MW Reference Wind Turbine," Energies, vol. 15, no. 20, 2022, https://doi.org/10.3390/en15207649.

17 WIND ENERGY↗
DOE OSTI2023

Probing individual single atom electrocatalyst sites by advanced analytical scanning transmission electron microscopy

Single atom electrocatalysts (SAEs) are promising next-generation materials for promoting a variety of important reactions, such as the oxygen reduction, nitrogen reduction, and CO 2 reduction reactions. While bulk characterization techniques such as X-ray absorption spectroscopy and Mössbauer spectroscopy have significantly enhanced our understanding of these catalysts, direct probing of individual single metal atom sites at the atomic scale is necessary to understand local variations in the properties of these sites and accelerate design and synthesis of improved SAEs. Aberration-corrected scanning transmission electron microscopy (STEM) has become a powerful tool for providing this type of atomic-scale information about SAE metal sites. These sites are typically unstable under the electron beam, however, which, in combination with conventional acquisition methods and detectors, has limited the type and quantity of information obtainable by spectroscopic STEM techniques. Here, we map multiple individual SAE metal sites in a nitrogen-doped carbon containing atomically dispersed Fe and Re (FeReNC) at the atomic scale by direct electron detection electron energy-loss spectroscopy (EELS). Direct electron detection provides an improved signal-to-noise ratio over conventional scintillator-based detectors and enables detection and real space localization of weak signals. In addition, we demonstrate an automated method for identification of metal atom positions, placement of the probe on these sites, and simultaneous EELS and energy dispersive X-ray spectroscopic (EDS) signal acquisition. This simultaneous acquisition of EELS and EDS provides access to the composition and bonding of a wide range of SAE metal sites. In this study, focusing the probe directly on the metal sites also increases the relevant data acquisition rate by more than an order of magnitude over two-dimensional mapping, enabling improved statistical measurements of site properties. The versatility, sensitivity, and speed that these techniques provide enhances our ability to probe the local elemental and chemical environment of a large number of individual SAE metal site structures at the atomic scale, enabling an improved understanding of the variations in the local properties of these electrocatalysts to be gained. As a result, significantly increased information about individual metal sites will be available to future electrochemical studies through these techniques, accelerating the development of advanced SAEs.

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

Improved Accuracy in Semi-Experimental Structure Determination by Resolving Problems Associated with Rotation of Principal Inertial Axes of Isotopologues: Structures of 1,3-Oxazole ( c -C 3 H 3 NO)

The rotational spectrum of the normal isotopologue of 1,3-oxazole (c-C 3 H 3 NO) was observed from 43 to 750 GHz. Over 3900 transitions for the ground vibrational state are measured, assigned, and least-squares fit to sextic centrifugally distorted-rotor Hamiltonians. The measured frequencies and resulting spectroscopic constants from this extended spectral range, combined with previous measurements of the nuclear quadrupole coupling constants, will facilitate astronomical searches for oxazole across the majority of the range of modern radiotelescopes. Spectra for a set of 30 oxazole isotopologues, which include multiple isotopic substitutions of each atom, are used to determine the first semi-experimental equilibrium ($r$$^{SE}_{e}$) structure and semi-experimental substitution structure ($r$$^{SE}_{e}$), each using CCSD(T) computed values for the vibration–rotation interaction and electron-mass corrections. The large number of isotopologues, including 21 isotopologues observed for the first time, and the redundant substitutions of each atom provide sufficient spectroscopic information to determine the $r$$^{SE}_{e}$ structure with the expected high level of accuracy and precision (0.0001 or 0.0002 Å in bond distances and 0.013 to 0.025° in bond angles). In the course of this study, we analyzed a known issue for some $r$$^{SE}_{e}$ structure determinations of near-oblate asymmetric tops in which inclusion of individual isotopologues degrades the structure determination. We demonstrate that this problem primarily arises from the difference in the values of the computed vibration–rotation interaction corrections as evaluated at the computed re geometry vs the $r$$^{SE}_{e}$ geometry of the “real” molecule. Our solution to this problem substantially improves the $r$$^{SE}_{e}$ structure of oxazole and likely can be generalized to many other molecules.

Chemical structure↗
DOE OSTI2022

DIII-D Electron Cyclotron Heating and Current Drive System Status and Plans

The DIII-D 110 GHz Electron Cyclotron Heating and Current Drive (ECH/ECCD) System supports a wide range of experiments at the DIII-D National Fusion Facility and is pioneering innovative new techniques for future fusion reactors. Here, the system currently includes 8 sockets with 4 installed gyrotrons, 31.75 mm circular corrugated transmission lines, four dual real time steerable launchers plus two Top Launchers. After the successful testing of the first Top Launcher, that injects the EC waves downward nearly parallel to the resonance layer with double current drive efficiency and a long absorption path, a second new Top Launcher was designed and installed with an upgraded design. Some of the operational limitations encountered for the 110 GHz X-mode second harmonic resonance scenario launched in over-dense plasmas were overcome with the alternative use of O-mode second harmonic launch. Two new gyrotrons in a new series manufactured by CPI were installed and tested, bringing up the total installed power to 3 MW. The installation of repaired and new gyrotrons will increase the number of installed gyrotrons to 7 in 2023. An addition to the DIII-D building will be completed in early 2022 to provide the space to expand the ECH/ECCD system to ten gyrotron systems, plus an extra gyrotron test socket.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Improve Learning from Crowds via Generative Augmentation

Crowdsourcing provides an efficient label collection schema for supervised machine learning. However, to control annotation cost, each instance in the crowdsourced data is typically annotated by a small number of annotators. This creates a sparsity issue and limits the quality of machine learning models trained on such data. In this paper, we study how to handle sparsity in crowdsourced data using data augmentation. Specifically, we propose to directly learn a classifier by augmenting the raw sparse annotations. We implement two principles of high-quality augmentation using Generative Adversarial Networks: 1) the generated annotations should follow the distribution of authentic ones, which is measured by a discriminator; 2) the generated annotations should have high mutual information with the ground-truth labels, which is measured by an auxiliary network. Extensive experiments and comparisons against an array of state-of-the-art learning from crowds methods on three real-world datasets proved the effectiveness of our data augmentation framework. It shows the potential of our algorithm for low-budget crowdsourcing in general.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗
DOE OSTI2020

TriC: Distributed-memory Triangle Counting by Exploiting the Graph Structure

Graph analytics has emerged as an important tool in the analysis of large scale data from diverse application domains such as social networks, cyber security and bioinformatics. Counting the number of triangles in a graph is a fundamental kernel with several applications such as detecting the community structure of a graph or in identifying important vertices in a graph. The ubiquity of massive datasets is driving the need to scale graph analytics on parallel systems. However, numerous challenges exist in efficiently parallelizing graph algorithms, especially on distributed-memory systems. Irregular memory accesses and communication patterns, low computation to communication ratios, and the need for frequent synchronization are some of the leading challenges. In this paper, we present TriC, our distributed-memory implementation of triangle counting in graphs using the Message Passing Interface (MPI), as a submission to the 2020 GraphChallenge competition. Using a set of synthetic and real-world inputs from the challenge, we demonstrate a speedup of up to 90x relative to previous work on 32 processor-cores of a NERSC Cori node. We also provide details from distributed runs with up to8192 processes along with strong scaling results. The observations presented in this work provide an understanding of the system-level bottlenecks at scale that specifically impact sparse-irregular workloads and will therefore benefit other efforts to parallelize graph algorithms.

Halappanavar, Mahantesh↗
DOE OSTI2024

Composite Qdrift-product formulas for quantum and classical simulations in real and imaginary time

Recent study has shown that it can be advantageous to implement a composite channel that partitions the Hamiltonian H for a given simulation problem into subsets A and B such that H = A + B , where the terms in A are simulated with a Trotter-Suzuki channel and the B terms are randomly sampled via the Qdrift algorithm. Here we extend Qdrift and composite product formulas to imaginary time, formulating candidate classical algorithms for quantum Monte Carlo calculations. We upper bound the induced Schatten- 1 → 1 norm on both imaginary-time Qdrift and composite channels. Another recent result demonstrated that simulations of lattice Hamiltonians containing geometrically local interactions can be improved using a Lieb-Robinson argument to decompose H into subsets that contain only terms supported on that subset of the lattice. Here, we provide a quantum algorithm by unifying this result with the composite approach into “local composite channels” and we upper bound the diamond distance. We provide exact numerical simulations of algorithmic cost by counting the number of gates of the form e − i H j t and e − H j β to meet a certain error tolerance ε . In doing so, we optimize the partitioning into sets A and B using gradient boosted tree models from machine learning. These numerical studies are important given that product formulas have been historically known to outperform analytic upper bounds. We show constant factor advantages for a variety of interesting Hamiltonians, the maximum of which is a ≈ 20 -fold speedup that occurs in the simulation of Jellium. Published by the American Physical Society 2024

Pocrnic, Matthew (ORCID:0000000203089376)↗