Finding the Number of Latent Topics With Semantic Non-Negative Matrix Factorization
Not Available
Vangara, Raviteja↗
SEARCH · Engineering Papers
Results for “Findings”
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.
An Active Learning-Based Streaming Pipeline for Reduced Data Training of Structure Finding Models in Neutron Diffractometry
Structure determination workloads in neutron diffractometry are computationally expensive and routinely require several hours to many days to determine the structure of a material from its neutron diffraction patterns. The potential for machine learning models trained on simulated neutron scattering patterns to significantly speed up these tasks have been reported recently. However, the amount of simulated data needed to train these models grows exponentially with the number of structural parameters to be predicted and poses a significant computational challenge. To overcome this challenge, we introduce a novel batch-mode active learning (AL) policy that uses uncertainty sampling to simulate training data drawn from a probability distribution that prefers labelled examples about which the model is least certain. We confirm its efficacy in training the same models with ∼ 75% less training data while improving the accuracy. We then discuss the design of an efficient stream-based training workflow that uses this AL policy and present a performance study on two heterogeneous platforms to demonstrate that, compared with a conventional training workflow, the streaming workflow delivers ∼ 20% shorter training time without any loss of accuracy.
Wang, Tianle [Brookhaven National Laboratory (BNL)↗
Finding Inputs that Trigger Floating-Point Exceptions in GPUs via Bayesian Optimization
Explore the source record for details and available documents.
Computer science↗
A Neural Network Approach for High-Dimensional Optimal Control Applied to Multiagent Path Finding
Not provided.
Automation & Control Systems↗
Building Trust in Earth Science Findings through Data Traceability and Results Explainability
Not Available
Olaya, Paula↗
Occasions for laughter and dementia risk: Findings from a six‐year cohort study
Aim Currently, there is little evidence on the relationship between laughter and the risk of dementia, and since laughter is mainly a social behavior, we aimed to examine the association between various occasions for laughter and the risk of dementia in Japanese older adults. Methods We draw upon 6‐year follow‐up data from the Japan Gerontological Evaluation Study, including 12 165 independent older adults aged 65 years or over. Occasions for laughter were assessed using a questionnaire, while dementia was diagnosed using the standardized dementia scale of the long‐term care insurance system in Japan. Cox proportional hazards models were estimated, yielding hazard ratios and 95% confidence intervals (CIs). Results The multivariable hazard ratio of dementia incidence for all participants in the groups for high versus low variety of occasions for laughter was 0.84 (95% CI: 0.72–0.98, P for trend <0.001). A greater variety of occasions for laughter was associated with a lower risk of dementia 0.78 (95% CI: 0.63–0.96, P for trend <0.001) among women, but was less pronounced for men, with significant associations only for the medium group. Laughing during conversations with friends, communicating with children or grandchildren, and listening to the radio were primarily associated with decreased risk. Conclusion A greater variety of laughter occasions in individual and social settings was associated with a reduced risk of dementia. Geriatr Gerontol Int 2022; 22: 392–398 .
Wang, Yu↗
Finding new friends and revisiting old ones – how plant lipid droplets connect with other subcellular structures
Summary The number of described contact sites between different subcellular compartments and structures in eukaryotic cells has increased dramatically in recent years and, as such, has substantially reinforced the well‐known premise that these kinds of connections are essential for overall cellular organization and the proper functioning of cellular metabolic and signaling pathways. Here, we discuss contact sites involving plant lipid droplets (LDs), including LD‐endoplasmic reticulum (ER) connections that mediate the biogenesis of new LDs at the ER, LD‐peroxisome connections, that facilitate the degradation of LD‐stored triacylglycerols (TAGs), and the more recently discovered LD‐plasma membrane connections, which involve at least three novel proteins, but have a yet unknown physiological function(s).
60 APPLIED LIFE SCIENCES↗
Finding MIDDLE Ground: Scalable and Secure Distributed Learning
Edge computing methods allow devices to efficiently train a high-performing, robust, and personalized model for predictive tasks. However, these methods succumb to privacy and scalability concerns such as adversarial data recovery and expensive model communication. Furthermore, edge computing methods unrealistically assume that all devices train an identical model. In practice, edge devices have varying computational and memory constraints which may not allow certain devices to have the space or speed to train a specific model. To overcome these issues, we propose MIDDLE: a model independent distributed learning algorithm which allows heterogeneous edge devices to assist each other’s training while communicating only non-sensitive information. MIDDLE unlocks the ability for edge devices, regardless of computational or memory constraints, to assist each other even with completely different model architectures. Furthermore, MIDDLE does not require model or gradient communication which greatly reduces communication size and time. We prove that MIDDLE attains the optimal convergence rate O(1/sqrt(TM)) of stochastic gradient descent for convex and non-convex smooth optimization (for total iterations T and batch size M). Finally, our experimental results demonstrate that MIDDLE (even in non-IID data settings) attains robust and high-performing models without model or gradient communication.
Bornstein, Marc I.↗
fpm-find v0.1.0
This program searches the fortran-lang community's package index (https://github.com/fortran-lang/webpage/blob/main/data/package_index.yml). The program is designed for use as a Fortran Package Manager (fpm) plugin.
Rouson, Damian [Lawrence Berkeley National Laborat↗
Neural Networks to Find the Optimal Forcing for Offsetting the Anthropogenic Climate Change Effects
Abstract Of great relevance to climate engineering is the systematic relationship between the radiative forcing to the climate system and the response of the system, a relationship often represented by the linear response function (LRF) of the system. However, estimating the LRF often becomes an ill-posed inverse problem due to high-dimensionality and nonunique relationships between the forcing and response. Recent advances in machine learning make it possible to address the ill-posed inverse problem through regularization and sparse system fitting. Here, we develop a convolutional neural network (CNN) for regularized inversion. The CNN is trained using the surface temperature responses from a set of Green’s function perturbation experiments as imagery input data together with data sample densification. The resulting CNN model can infer the forcing pattern responsible for the temperature response from out-of-sample forcing scenarios. This promising proof of concept suggests a possible strategy for estimating the optimal forcing to negate certain undesirable effects of climate change. The limited success of this effort underscores the challenges of solving an inverse problem for a climate system with inherent nonlinearity. Significance Statement Predicting the climate response for a given climate forcing is a direct problem, while inferring the forcing for a given desired climate response is often an inverse, ill-posed, problem, posing a new challenge to the climate community. This study makes the first attempt to infer the radiative forcing for a given target pattern of global surface temperature response using a deep learning approach. The resulting deeply trained convolutional neural network inversion model shows promise in capturing the forcing pattern corresponding to a given surface temperature response, with a significant implication on the design of an optimal solar radiation management strategy for curbing global warming. This study also highlights the technical challenges that future research should prioritize in seeking feasible solutions to the inverse climate problem.
Ren, Huiying↗
Finding the perfect imperfection: Accelerated, computationally driven discovery and design of quantum defects
Optically addressable spin defects have emerged as the leading platforms for quantum sensing and communication in solid-state systems. While traditional efforts have concentrated on a focused set of well-studied defects, recent advances in high-throughput computational methods have shown promise for large-scale exploration of defects across diverse semiconductor hosts. By cataloging key properties of quantum defects in computational databases, high-throughput screening techniques can systematically suggest and design novel candidates. In this article, we highlight recent advances in data-driven quantum defect design aimed at addressing critical materials science challenges such as host materials selection, defect stability, and desirable electronic and optical properties. Here, we emphasize the importance of electronic-structure-guided searches across various materials and illustrate how high-throughput computations contribute to our understanding of design principles for quantum defects. Additionally, we outline ongoing challenges and emerging opportunities in this rapidly developing field.
Xiong, Yihuang [Dartmouth College, Hanover, NH (Un↗
Modernist materials synthesis: Finding thermodynamic shortcuts with hyperdimensional chemistry
Synthesis remains a challenge for advancing materials science. A key focus of this challenge is how to enable selective synthesis, particularly as it pertains to metastable materials. This perspective addresses the question: how can “spectator” elements, such as those found in double ion exchange (metathesis) reactions, enable selective materials synthesis? By observing reaction pathways as they happen (in situ) and calculating their energetics using modern computational thermodynamics, we observe transient, crystalline intermediates that suggest that many reactions attain a local thermodynamic equilibrium dictated by local chemical potentials far before achieving a global equilibrium set by the average composition. Furthermore, using this knowledge, one can thermodynamically “shortcut” unfavorable intermediates by including additional elements beyond those of the desired target, providing access to a greater number of intermediates with advantageous energetics and selective phase nucleation. Ultimately, data-driven modeling that unites first-principles approaches with experimental insights will refine the accuracy of emerging predictive retrosynthetic models for complex materials synthesis.
36 MATERIALS SCIENCE↗
Improving Discovery, Sharing, and Use of Water Data: Initial Findings and Suggested Future Work
Collaborative management of water resources requires a broad suite of “water data” that extends beyond basic information about water quantity and quality to other related topics such as water infrastructure, aquatic ecosystem health, socioeconomic factors, and power generation. Water data are disparate in nature because they are collected and provided by many entities, and in some cases, remain challenging to access and use. The U.S. Department of Energy’s Water Power Technologies Office initiated a project to characterize relevant categories of water data; describe the current state of accessing, using, and visualizing water data; and outline investigatory pathways for future efforts aimed at improving the discovery, sharing, and use of water data. Input on these topics was solicited from a small but diverse cross section of members of the water resources community. Fourteen broad categories of water data were identified: dams; ecology; flood control; hydroclimatology; hydrography; hydrology; hydropower; management landscape; migratory barriers; recreation and aesthetic importance; socioeconomic; water quality; water availability and use; and weather. Stakeholder perspectives on the accessibility and usability of water data indicate these aspects are affected by a complex set of technical and social factors. However, stakeholders generally agreed that better access to water data can provide a range of benefits to water management, and they stressed the need to generate broad support from water data users and producers. Two investigatory pathways were outlined that, taken together, provide a logical progression toward the goals of the project. The first pathway emphasizes further investigation to better define target audiences and data needs, identify opportunities for collaboration between related efforts, and conduct value demonstration activities to generate further support for improving discovery and access of water data. The second pathway focuses on creating a comprehensive vision for potential solutions that improve the discovery of water data. Several activities that align with the first pathway are suggested for the next phase of the project.
13 HYDRO ENERGY↗
Interim Findings and Suggestions: Cyclic Steaming above Fracturing Pressure and the Associated Surface Expressions
In the wake of California’s new regulations for Underground Injection Control (enacted in April 2019) and several recent high-visibility surface expressions associated with high-pressure cyclic steaming in diatomite formations, a Moratorium on the issuance of permits for new wells using high pressure steam injection was announced on November 19, 2019. We, a panel consisting of technical experts and regulators from CalGEM and Lawrence Livermore National Laboratory, were commissioned by CalGEM to carry out the study as documented in this report, to: Determine if projects using cyclic steam injection above fracturing pressure (referred to as “cyclic steaming” hereafter) can be implemented safely, and Identify specific criteria for future permitting should the Moratorium be lifted.
02 PETROLEUM↗
Deep learning to help find continuous gravitational waves [Slides]
This presentation is a celebration of research that was mentored by Grant David Meadors during the summer. The hard work was done by the students at the ICR in Santa Fe, July 2021.
79 ASTRONOMY AND ASTROPHYSICS↗
Report Series: Evaluation, Finding of Effect, and Mitigation Documentation for the Main Gate (23-GS100), Mercury, Area 23, Nevada National Security Site, Nye County, Nevada
The U.S. Department of Energy (DOE) National Nuclear Security Administration Nevada Field Office (NNSA/NFO) plans to replace the existing guard shack at the Main Gate (23-GS100, Nevada State Historic Preservation Office [SHPO] Resource No. S1758) to the Nevada National Security Site (NNSS) in Nye County, Nevada. The purpose of the project is to improve security. The project is considered an undertaking subject to review under Title 54 of United States Code (USC) § 306108, commonly known as Section 106 of the National Historic Preservation Act, Title 54 USC § 300101, et seq., and its implementing regulations, Title 36 of the Code of Federal Regulations (36 CFR) Part 800. In 2018, Desert Research Institute (DRI) completed an architectural survey of the town of Mercury. This effort resulted in the identification, recordation, and evaluation of the Mercury Historic District (MHD, SHPO Resource No. D230), including the identification of its contributing elements (Reno et al. 2018). The MHD was recommended eligible for listing in the National Register of Historic Places (NRHP, National Register) under the Secretary of the Interior’s (SOI) Significance Criteria A and C, as defined in 36 CFR Part 60.4, as a significant concentration of buildings and structures with a direct and important association with Cold War-era nuclear testing from 1951 through 1992. It has not been evaluated under Criteria B and D to date. As part of a larger modernization program for Mercury, the NNSA/NFO and the SHPO executed the 2018 Programmatic Agreement between the National Nuclear Security Administration Nevada Field Office and the Nevada State Historic Preservation Officer regarding Modernization and Operational Maintenance of the Nevada National Security Site, at Mercury in Nye County, Nevada (Mercury PA). The Mercury PA includes streamlined Section 106 procedures for undertakings in the MHD based on contributing element categories. The Main Gate is identified in Appendix C of the Mercury PA as a Category I contributing element, indicating that it might be individually eligible for the NRHP. It is a historic property for the purposes of Section 106 compliance and subject to the stipulations of the Mercury PA. Per Stipulation VI of the Mercury PA, when the Area of Potential Effect (APE) for an undertaking includes Category I elements, the NNSA/NFO must evaluate the Category I elements for individual NRHP eligibility under all of the SOI Significance Criteria, prior to initiating any activity that may affect the elements. Thus, the purpose of this report is to evaluate the Main Gate as a potential individually eligible historic property in fulfillment of Stipulation VI of the Mercury PA. The evaluation detailed herein concludes that the Main Gate is individually eligible for listing in the NRHP under SOI Significance Criterion A at the national level of significance for its direct, important association with Cold War-era nuclear testing from 1965 (the date the current Main Gate was constructed) through 1992 (when critical nuclear testing on the NNSS ceased).
23-GS100↗