Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “topic modeling”

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 55 records · Page 3

Distributed-Memory Parallel JointNMF

Joint Nonnegative Matrix Factorization (JointNMF) is a hybrid method for mining information from datasets that contain both feature and connection information. We propose distributed-memory parallelizations of three algorithms for solving the JointNMF problem based on Alternating Nonnegative Least Squares, Projected Gradient Descent, and Projected Gauss-Newton. We extend well-known communication-avoiding algorithms using a single processor grid case to our coupled case on two processor grids. We demonstrate the scalability of the algorithms on up to 960 cores (40 nodes) with 60% parallel efficiency. The more sophisticated Alternating Nonnegative Least Squares (ANLS) and Gauss-Newton variants outperform the first-order gradient descent method in reducing the objective on large-scale problems. We perform a topic modelling task on a large corpus of academic papers that consists of over 37 million paper abstracts and nearly a billion citation relationships, demonstrating the utility and scalability of the methods.

Eswar, Srinivas↗

HPC-NMF: A High-Performance Parallel Algorithm for Nonnegative Matrix Factorization

NMF is a useful tool for many applications in different domains such as topic modeling in text mining, background separation in video analysis, and community detection in social networks. Despite its popularity in the data mining community, there is a lack of efficient distributed algorithms to solve the problem for big data sets. We propose a high-performance distributed-memory parallel algorithm that computes the factorization by iteratively solving alternating non-negative least squares (NLS) subproblems for $\WW$ and $\HH$. It maintains the data and factor matrices in memory (distributed across processors), uses MPI for interprocessor communication, and, in the dense case, provably minimizes communication costs (under mild assumptions). As opposed to previous implementation, our algorithm is also flexible: It performs well for both dense and sparse matrices, and allows the user to choose any one of the multiple algorithms for solving the updates to low rank factors $\WW$ and $\HH$ within the alternating iterations.

Kannan, Ramakrishnan↗

gallimgg/Topic_Modeling_Assistant

Topic Modeling Assistant for large literature databases in Scopus system with zettelkasten note exporting tool for use in Obsidian

Gallimore, GrantGregory [Oak Ridge National Labora↗

ESTEEM: A Novel Framework for Qualitatively Evaluating and Visualizing Spatiotemporal Embeddings in Social Media

Analyzing and visualizing large amounts of social media communications and contrasting short-term conversation changes over time and geo-locations is extremely important for commercial and government applications. Earlier approaches for large-scale text stream summarization used dynamic topic models and trending words. Instead, we rely on text embeddings – low-dimensional word representations in a continuous vector space where similar words are embedded nearby each other. This paper presents ESTEEM,1 a novel tool for visualizing and evaluating spatiotemporal embeddings learned from streaming social media texts. Our tool allows users to monitor and analyze query words and their closest neighbors with an interactive interface. We used state-of- the-art techniques to learn embeddings and developed a visualization to represent dynamically changing relations between words in social media over time and other dimensions. This is the first interactive visualization of streaming text representations learned from social media texts that also allows users to contrast differences across multiple dimensions of the data.

social media dynamics↗

Report in the Energy and Intensity Frontiers, and Theoretical at Northwestern University

The Northwestern (NU) Particle Physics (PP) group involved in this report is active on all the following priority areas: Energy and Intensity Frontiers. The group is lead by 2 full profs. in experimental physics (Schmitt and Velasco), 3 full profs. in theoretical physics (de Gouvea, Low and Petriello), and Heidi Schellman who is now at Oregon State. Low and Petriello hold joint appointments with the HEP Division at Argonne National Laboratory. The theoretical PP research focuses on different aspects of PP phenomenology. de Gouvea dedicates a large fraction of his research efforts to understanding the origin of neutrino masses, neutrino properties and uncovering other new phenomena, and investigating connections between neutrino physics and other aspects of PP. Low works on Higgs physics as well as new theories beyond the Standard Model. Petriello pursues a research program in precision QCD and its associated collider phenomenology. The main goal of this effort is to improve the Standard Model predictions for important LHC observables in order to enable discoveries of new physics. In recent years, the emphasis on experimental PP at NU has been in collider physics. NU expands its efforts in new directions in both the Intensity and the Cosmic Frontiers (not discussed in this report). In the Intensity Frontier, Schmitt has started a new effort on Mu2e. He was accepted as a collaborator in April 2015 and is identified with important projects. In the Energy Frontier, Hahn, Schmitt and Velasco continue to have a significant impact and expanded their CMS program to include R&D for the real-time L1 tracking trigger and the high granularity calorimeter needed for the high-luminosity LHC. Hahn is supported by an independent DOE Career Award and his work will not be discussed in this document. The NU analysis effort includes searches for rare and forbidden decays of the Higgs bosons, Z boson, top quark, dark matter and other physics beyond the standard model topics. Four students completed their PhD: Kubik is now contributing to the Cosmic Frontier program, Pollack to both the Intensity and Energy Frontiers and Pozdnyakov and Odell will continue in the Energy Frontier. All our research scientists, Anastassov, Oferzynski, Lusito, and Stoynev, have found new positions. The new post-docs are Trovato from Scuola Normale de Pisa, Odell from Northwestern and Bhattacharya from Brown. Trovato is now supported by Hahn, and so is Sung, previously at MIT.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Deep Underground Neutrino Experiment (DUNE) Near Detector Conceptual Design Report

The Deep Underground Neutrino Experiment (DUNE) is an international, world-class experiment aimed at exploring fundamental questions about the universe that are at the forefront of astrophysics and particle physics research. DUNE will study questions pertaining to the preponderance of matter over antimatter in the early universe, the dynamics of supernovae, the subtleties of neutrino interaction physics, and a number of beyond the Standard Model topics accessible in a powerful neutrino beam. A critical component of the DUNE physics program involves the study of changes in a powerful beam of neutrinos, i.e., neutrino oscillations, as the neutrinos propagate a long distance. The experiment consists of a near detector, sited close to the source of the beam, and a far detector, sited along the beam at a large distance. This document, the DUNE Near Detector Conceptual Design Report (CDR), describes the design of the DUNE near detector and the science program that drives the design and technology choices. The goals and requirements underlying the design, along with projected performance are given. It serves as a starting point for a more detailed design that will be described in future documents.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Machine Learning for Identifying Relevance to Biosurveillance in Multilingual Text

Objective: The objective is to develop an ensemble of machine learning algorithms to identify multilingual, online articles that are relevant to biosurveillance. Language morphology varies widely across languages and must be accounted for when designing algorithms. Here, we compare the performance of a word embedding-based approach and a topic modeling approach with machine learning algorithms to determine the best method for Chinese, Arabic, and French languages. Introduction: Global biosurveillance is an extremely important, yet challenging task. One form of global biosurveillance comes from harvesting open source online data (e.g. news, blogs, reports, RSS feeds). The information derived from this data can be used for timely detection and identification of biological threats all over the world. However, the more inclusive the data harvesting procedure is to ensure that all potentially relevant articles are collected, the more data that is irrelevant also gets harvested. This issue can become even more complex when the online data is in a non-native language. Foreign language articles not only create language-specific issues for Natural Language Processing (NLP), but also add a significant amount of translation costs. Previous work shows success in the use of combinatory monolingual classifiers in specific applications, e.g., legal domain [1]. A critical component for a comprehensive, online harvesting biosurveillance system is the capability to identify relevant foreign language articles from irrelevant ones based on the initial article information collected, without the additional cost of full text retrieval and translation.

Machine learning↗

Advanced Analytics Studies Applied to US Department of Veterans Affairs' Corporate Data Warehouse (Initial Draft)

In this report, we describe our experience of applying several advanced analytics algorithms to the US Department of Veterans Affairs’ (VA’s) Corporate Data Warehouse (CDW) electronic health records datasets during FY2017-18. While various algorithms were applied to the CDW data, the main goal of this effort was to provide useful insights into the operational aspects (as opposed to the purely clinical aspect) of this specific implementation. Since there are not many reports in the public literature on advanced analytics applied to this data, this report provides unique insight in this regard. We focused on two machine learning (ML) applications: (1) clinical pathway inference (patterns dominating clinical pathways applied to a male dataset of Stable Ischemic Heart Disease patients as use case) and (2) medical-concept representation learning’s capability to inform and refine cohort membership based on probable patient outcomes. During our study on clinical pathway inference, we used two main modeling techniques: (a) topic probabilistic modeling (latent Dirichlet allocation [LDA]) and (b) feature imbedding. We observed that the applicability of LDA to the pathway inference remains in question. LDA provides results that are intuitive and easy for humans to comprehend; however, when applied to the pathway inference, it also generates pathway patterns that consist of a few dominant pathway components. In addition, LDA provides no information regarding the relationships among components in the same pathway. The outputs of LDA are also found to categorize patients based on their trace data of clinical procedures. However, the results suggest that LDA is biased toward statistically dominant components. This makes it particularly hard to discriminate pathway subbranches. To address the issue, we designed a new pathway inference methodology that integrates temporal ordering of pathway components into LDA results, as existing feature embedding tools such as word2vec are not readily applicable to our task. In addition, we developed our own feature embedding tool customized for patient traces. We conclude this study by suggesting two approaches to use embedding representation of a component into LDA and briefly list our lessons learned and recommendations for future work. For our study on medical-concept representation learning ability to inform and refine cohort membership based on probable patient outcomes, during our empirical evaluation we compared methods from medical-concept learning to standard one of a kind (one-of-K) encoding to evaluate the change in effectiveness as done in Choi, Schuetz, Stewart, and Sun (2016). We performed two primary empirical evaluations. The first evaluation was on a curated collection of 60,000 patients with no more than 1 year of medical history included. The second was a collection of patients with no restriction to the amount of medical history included. The hypothesis across both evaluations is that representation learning is broadly useful independent of downstream processing models; thus, for each experiment, we trained three models—a logistic regression model, a two-layer neural network model, and a nearest-neighbor model—and then averaged the results. Each model was trained using fivefold cross validation. We provided evidence that medical-representation learning improves predictions of the primary diagnosis category of a short patient history through 12 experiments. In addition, we provided evidence that medical-representation learning fails to improve prediction of the primary diagnosis category of an arbitrarily long patient history through 12 experiments. We conclude this study with a brief list of our lessons learned and recommendations for future work. We conclude the report with a discussion of recent technology trends as they relate to our artificial intelligence and, more specifically, to our ML research and approaches as described previously.

97 MATHEMATICS AND COMPUTING↗

The Role of Chemistry in Fracture Pattern Development and Opportunities to Advance Interpretations of Geological Materials

Abstract Fracture pattern development has been a challenging area of research in the Earth sciences for more than 100 years. Much has been learned about the spatial and temporal complexity inherent to these systems, but severe challenges remain. Future advances will require new approaches. Chemical processes play a larger role in opening‐mode fracture pattern development than has hitherto been appreciated. This review examines relationships between mechanical and geochemical processes that influence the fracture patterns recorded in natural settings. For fractures formed in diagenetic settings (~50 to 200 °C), we review evidence of chemical reactions in fractures and show how a chemical perspective helps solve problems in fracture analysis. We also outline impediments to subsurface pattern measurement and interpretation, assess implications of discoveries in fracture history reconstruction for process‐based models, review models of fracture cementation and chemically assisted fracture growth, and discuss promising paths for future work. To accurately predict the mechanical and fluid flow properties of fracture systems, a processes‐based approach is needed. Progress is possible using observational, experimental, and modeling approaches that view fracture patterns and properties as the result of coupled mechanical and chemical processes. A critical area is reconstructing patterns through time. Such data sets are essential for developing and testing predictive models. Other topics that need work include models of crystal growth and dissolution rates under geological conditions, cement mechanical effects, and subcritical crack propagation. Advances in machine learning and 3‐D imaging present opportunities for a mechanistic understanding of fracture formation and development, enabling prediction of spatial and temporal complexity over geologic timescales. Geophysical research with a chemical perspective is needed to correctly identify and interpret fractures from geophysical measurements during site characterization and monitoring of subsurface engineering activities.

58 GEOSCIENCES↗

Natural Language Processing to Inform Agent-Based Modeling: With Application to Modeling Adoption of Medium-Duty Electric Vehicles

Agent-based socio-technical modeling of medium- and heavy-duty (MDHD) electric vehicle (EV) adoption has the potential to provide analysis, prediction, and gui. This paper describes new applications of text analysis developed through machine learning (ML) to build and understand relevant topics and their saliency in the published discourse on adoption of MDHD EVs. This work contributes to the state of the art in topic mining models by defining a new metric of topic ranking (START) that quantifies the importance of predefined topics within the corpus using weighted results for predefined topics from two topic modeling approaches: Latent Dirichlet Allocation (LDA) and BERTopic. The START metric is then demonstrated in practice to model how academia and industry view the EV adoption process based on the respective texts published by these groups. Results show that academic literature places more emphasis on categories of interests such as norms/attitudes and adopter knowledge, while trade journals tend to emphasize long-term cost more than academia. The two bodies of literature agree on the importance of policy and incentives in MDHD EV adoption. Together these results illustrate the potential to use ML-based text analysis to populate the characteristics of agent-based socio-technical models.

Electric vehicle adoption, fleet electrification, ↗

Topical Report on Water Balance Model Development and Initial Application

This topical report summarizes results from developing a generic dynamic water balance tool which can be used by coal-fired power plant operators to evaluate FGD water management strategies. EPRI is using a statistical software model created on the GoldSim platform and data from coal-fired energy plants to determine the fraction of power plant water needs that can be satisfied by electrodialysis reversal treatment and, correspondingly, the discharge volume and/or byproducts that still need to be managed. The generic model is designed to accept a number of easily obtainable process variables and generate probabilities of resultant water characteristics. The model is also constructed as a universal process flow; one that can be adjusted to accommodate most major components of an FGD blowdown handling system. The generic model is dynamic and utilizes user-inputted operational variation in water quality and quantity to best provide a probabilistic representation of potential performance outcomes. In order to account for complexities of water composition and treatment, a separate chemical modeling software is used to evaluate water chemistry and unit operation performance. The thermodynamic modeling software, OLI Studio, is employed to convert ion data from laboratory analyses into balanced molecular concentrations that are classified as dissolved and suspended solids. The dynamic water model is designed to import a total of 11 species commonly found in FGD blowdown water as molecular concentrations.Water management unit operations are used in the model and categorized by either physical unit operation blocks or chemical unit operations blocks. As the name implies, the physical water treatment blocks are unit operations that are exclusively physical changes to the system (e.g., separations and volume changes). The chemical water treatment blocks are unit operations where chemical changes occur and require the use of OLI Studio to characterize subsequent streams. Water and material balances from three unique coal-fired power plants were obtained and preliminary modeling on the dynamic water balance tool was performed using the extracted data from these power plants. These three power plants were chosen due to their unique configurations and water profiles that represented a variety of operational changes. Results from using the dynamic tool with the three plant configurations are presented and discussed in the context of the existing tool and planned future capabilities.

01 COAL, LIGNITE, AND PEAT↗

Topical Report on Water Balance Model Development and Initial Application

This topical report summarizes results from developing a generic dynamic water balance tool which can be used by coal-fired power plant operators to evaluate FGD water management strategies. EPRI is using a statistical software model created on the GoldSim platform and data from coal-fired energy plants to determine the fraction of power plant water needs that can be satisfied by advanced treatment technologies and, correspondingly, the discharge volume and/or byproducts that still need to be managed. The initial evaluations presented in this report represent generic scenarios, while the final evaluations will incorporate results from advanced treatment and encapsulation studies conducted during later stages of the project.

01 COAL, LIGNITE, AND PEAT↗

Radar-Cross-Section Reduction of Wind Turbines (Part 1)

In recent years, increasing deployment of large wind-turbine farms has become an issue of growing concern for the radar community. The large radar cross section (RCS) presented by wind turbines interferes with radar operation, and the Doppler shift caused by blade rotation causes problems identifying and tracking moving targets. Each new wind-turbine farm installation must be carefully evaluated for potential disruption of radar operation for air defense, air traffic control, weather sensing, and other applications. Several approaches currently exist to minimize conflict between wind-turbine farms and radar installations, including procedural adjustments, radar upgrades, and proper choice of low-impact wind-farm sites, but each has problems with limited effectiveness or prohibitive cost. An alternative approach, heretofore not technically feasible, is to reduce the RCS of wind turbines to the extent that they can be installed near existing radar installations. This report summarizes efforts to reduce wind-turbine RCS, with a particular emphasis on the blades. The report begins with a survey of the wind-turbine RCS-reduction literature to establish a baseline for comparison. The following topics are then addressed: electromagnetic model development and validation, novel material development, integration into wind-turbine fabrication processes, integrated-absorber design, and wind-turbine RCS modeling. Related topics of interest, including alternative mitigation techniques (procedural, at-the-radar, etc.), an introduction to RCS and electromagnetic scattering, and RCS-reduction modeling techniques, can be found in a previous report.

17 WIND ENERGY↗

SIMPLIFIED PREDICTIVE MODELS FOR CO₂ SEQUESTRATION PERFORMANCE ASSESSMENT RESEARCH TOPICAL REPORT ON TASK #3 STATISTICAL LEARNING BASED MODELS

We compare two approaches for building a statistical proxy model (metamodel) for CO₂ geologic sequestration from the results of full-physics compositional simulations. The first approach involves a classical Box-Behnken or Augmented Pairs experimental design with a quadratic polynomial response surface. The second approach used a space-filling maxmin Latin Hypercube sampling or maximum entropy design with the choice of five different meta-modeling techniques: quadratic polynomial, kriging with constant and quadratic trend terms, multivariate adaptive regression spline (MARS) and additivity and variance stabilization (AVAS). Simulations results for CO₂ injection into a reservoir-caprock system with 9 design variables (and 97 samples) were used to generate the data for developing the proxy models. The fitted models were validated with using an independent data set and a cross-validation approach for three different performance metrics: total storage efficiency, CO₂ plume radius and average reservoir pressure. The Box-Behnken–quadratic polynomial metamodel performed the best, followed closely by the maximin LHS–kriging metamodel.

Mishra, Srikanta↗

Single crystal toroidal diamond anvils for high pressure experiments beyond 5 megabar

Static compression experiments over 4 Mbar are rare, yet critical for developing accurate fundamental physics and chemistry models, relevant to a range of topics including modeling planetary interiors. Here we show that focused ion beam crafted toroidal single-crystal diamond anvils with ~9.0 μm culets are capable of producing pressures over 5 Mbar. The toroidal surface prevents gasket outflow and provides a means to stabilize the central culet. We have reached a maximum pressure of ~6.15 Mbar using Re as in situ pressure marker, a pressure regime typically accessed only by double-stage diamond anvils and dynamic compression platforms. In conclusion, optimizing single-crystal diamond anvil design is key for extending the pressure range over which studies can be performed in the diamond anvil cell.

36 MATERIALS SCIENCE↗

Investigation of Ultrahigh-Pressure Phase Transitions in Metals with a Toroidal Diamond Anvil Cell

Static compression experiments over 4 Mbar are rare, yet critical for developing accurate fundamental physics and chemistry models, relevant to a range of topics including modeling planetary interiors. Within this project we developed a toroidal diamond anvil geometry, crafted into single crystal diamonds with focused ion beam, with 9 μm culet proved capable of achieving pressures in excess of 5 megabars. The toroidal surface prevents gasket outflow and provides a means to stabilize the central culet. We have reached a maximum pressure of ~6.15 Mbar using Re as in-situ pressure marker, a pressure regime typically accessed only by double-stage diamond anvils and dynamic compression platforms. Moreover, utilizing these anvils for studying various transition metals the designs has proven to be a reliable platform, enabling routine static high-pressure studies in the 3 to 5 megabar pressure regime.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗