Data-Driven Model for Fire Spread using Deep Neural Networks
Explore the source record for details and available documents.
SEARCH · Engineering Papers
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.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Meridional flow is thought to play a very important role in the dynamics of the solar convection zone; however, because of its relatively small amplitude, precisely measuring it poses a significant challenge. Here we present a complete time–distance helioseismic analysis of about 2 years of ground-based Global Oscillation Network Group (GONG) Doppler data to retrieve the meridional circulation profile for modest latitudes in an attempt to corroborate results from other studies. We use an empirical correction to the travel times due to an unknown center-to-limb systematic effect. The helioseismic inversion procedure is first tested and reasonably validated on artificial data from a large-scale numerical simulation followed by a test to broadly recover the solar differential rotation found from global seismology. From GONG data, we measure poleward photospheric flows at all latitudes with properties that are comparable with earlier studies and a shallow equatorward flow about 65 Mm beneath the surface, in agreement with recent findings from Helioseismic and Magnetic Imager (HMI) data. No strong evidence of multiple circulation cells in depth or latitude is found, yet the whole phase space has not yet been explored. Tests of mass flux conservation are then carried out on the inferred GONG and HMI flows and compared to a fiducial numerical baseline from models, and we find that the continuity equation is poorly satisfied. While the two disparate data sets do give similar results for about the outer 15% of the interior radius, the total inverted circulation pattern appears to be unphysical in terms of mass conservation when interpreted over modest time scales. We can likely attribute this to both the influence of realization noise and subtle effects in the data and measurement procedure.
We have conducted a search for strong gravitational lensing systems in the Dark Energy Spectroscopic Instrument (DESI) Legacy Imaging Surveys Data Release 9. This is the third paper in a series. These surveys together cover ~19,000 deg 2 visible from the Northern Hemisphere, reaching a z-band AB magnitude of ~22.5. We use a deep residual neural network, trained on a compilation of known lensing systems and high-grade candidates as well as nonlenses in the same footprint. After applying our trained neural network to the survey data, we visually inspect and rank images with probabilities above a threshold which has been chosen to balance precision and recall. We have found 1895 lens candidates, of which 1512 are identified for the first time. Combining the discoveries from this work with those from Papers I (335) and II (1210), we have discovered a total of 3057 new candidates in the Legacy Surveys.
The computational intensity of detector simulation and event reconstruction poses a significant difficulty for data analysis in collider experiments. This challenge inspires the continued development of machine learning techniques to serve as efficient surrogate models. We propose a fast emulation approach that combines simulation and reconstruction. In other words, a neural network generates a set of reconstructed objects conditioned on input particle sets. To make this possible, we advance set-conditional set generation with diffusion models. Using a realistic, generic, and public detector simulation and reconstruction package (COCOA), we show how diffusion models can accurately model the complex spectrum of reconstructed particles inside jets.
Wireless networking and mobile communications is increasing around the world and in all sectors of our lives. With increasing use, the density and complexity of the systems increase with more base stations and advanced protocols to enable higher data throughputs. The security of data transported over wireless networks must also evolve with the advances in technologies enabling more capable wireless networks. However, means for analysis of the effectiveness of security approaches and implementations used on wireless networks are lacking. More specifically a capability to analyze the lower-layer protocols (i.e., Link and Physical layers) is a major challenge. An analysis approach that incorporates protocol implementations without the need for RF emissions is necessary. In this research paper several emulation tools and custom extensions that enable an analysis platform to perform cyber security analysis of lower layer wireless networks is presented. A use case of a published exploit in the 802.11 (i.e., WiFi) protocol family is provided to demonstrate the effectiveness of the described emulation platform.
Complex Markov models are widely used and powerful predictive tools to analyze stochastic biochemical processes. However, when the network of states is unknown, it is necessary to extract information from the data to partially build the network and estimate the values of the rates. The short-time behavior of the first-passage time distributions between two states in linear chains has been shown recently to behave as a power of time with an exponent equal to the number of intermediate states. For a general Markov model we derive the complete Taylor expansion of the first-passage time distribution between two arbitrary states. By combining algebraic methods and graph theory approaches it is shown that the first term of the Taylor expansion is determined by the shortest path from the initial state to the final state. When this path is unique, we prove that the coefficient of the first term can be written in terms of the product of the transition rates along the path. It is argued that the application of our results to first-return times may be used to estimate the dependence of rates on external parameters in experimentally measured time distributions.
The various technologies presented herein relate to determining a network attack is taking place, and further to adjust one or more network parameters such that the network becomes dynamically configured. A plurality of machine learning algorithms are configured to recognize an active attack pattern. Notification of the attack can be generated, and knowledge gained from the detected attack pattern can be utilized to improve the knowledge of the algorithms to detect a subsequent attack vector(s). Further, network settings and application communications can be dynamically randomized, wherein artificial diversity converts control systems into moving targets that help mitigate the early reconnaissance stages of an attack. An attack(s) based upon a known static address(es) of a critical infrastructure network device(s) can be mitigated by the dynamic randomization. Network parameters that can be randomized include IP addresses, application port numbers, paths data packets navigate through the network, application randomization, etc.
The discovery toolset for Emulytics enables the construction of high-fidelity emulation models of systems. The toolset consists of a set of tools and techniques to automatically go from network discovery of operational systems to emulating those complex systems. Our toolset combines data from host discovery and network mapping tools into an intermediate representation that can then be further refined. Once the intermediate representation reaches the desired state, our toolset supports emitting the Emulytics models with varying levels of specificity based on experiment needs.
Herein, early prediction of battery performance degradation trends can facilitate research of new materials and cell designs, rapid deployment of batteries in real-world applications, timely replacement of batteries in critical applications, and even the secondary use market. In this study, we design a convolutional neural network model to predict the entire battery capacity fade curve - a critical indicator of battery performance degradation - using first 100 cycles of data (~ three weeks of testing). We use the discharge voltage-capacity curves as input to the model and automate the feature extraction process through the convolutional layers of the network. Our approach can predict the per cycle capacity fade rate and rollover cycle (knee point) in the capacity fade curve, which indicate the onset of rapid capacity decay. On the publicly available graphite/LiFePO 4 battery dataset, optimized networks predict the capacity fade curves, rollover cycle, and end of life with 3.7% (worst-case), 19%, and 17% mean absolute percentage errors, respectively.
We describe a novel method for the application of convolutional neural networks (CNNs) to fields defined on the sphere, using the Hierarchical Equal Area Latitude Pixelization scheme (HEALPix). Specifically, we have developed a pixel-based approach to implement convolutional and pooling layers on the spherical surface, similarly to what is commonly done for CNNs applied to Euclidean space. The main advantage of our algorithm is to be fully integrable with existing, highly optimized libraries for NNs (e.g., PyTorch, TensorFlow, etc.). We present two applications of our method: (i) recognition of handwritten digits projected on the sphere; (ii) estimation of cosmological parameter from simulated maps of the cosmic microwave background (CMB). The latter represents the main target of this exploratory work, whose goal is to show the applicability of our CNN to CMB parameter estimation. We have built a simple NN architecture, consisting of four convolutional and pooling layers, and we have used it for all the applications explored herein. Concerning the recognition of handwritten digits, our CNN reaches an accuracy of ~95%, comparable with other existing spherical CNNs, and this is true regardless of the position and orientation of the image on the sphere. For CMB-related applications, we tested the CNN on the estimation of a mock cosmological parameter, defining the angular scale at which the power spectrum of a Gaussian field projected on the sphere peaks. We estimated the value of this parameter directly from simulated maps, in several cases: temperature and polarization maps, presence of white noise, and partially covered maps. For temperature maps, the NN performances are comparable with those from standard spectrum-based Bayesian methods. For polarization, CNNs perform about a factor four worse than standard algorithms. Nonetheless, our results demonstrate, for the first time, that CNNs are able to extract information from polarization fields, both in full-sky and masked maps, and to distinguish between E and B -modes in pixel space. Lastly, we have applied our CNN to the estimation of the Thomson scattering optical depth at reionization ( τ ) from simulated CMB maps. Even without any specific optimization of the NN architecture, we reach an accuracy comparable with standard Bayesian methods. This work represents a first step towards the exploitation of NNs in CMB parameter estimation and demonstrates the feasibility of our approach.
Large radio telescopes, cyber-security systems monitoring real-time network traffic, and others have specialized data storage needs: guaranteed capture of an ultra-high-bandwidth data stream, retention of the data long enough to determine what is 'interesting,' retention of interesting data indefinitely, and concurrent read/write access to determine what data is interesting, without interrupting the ongoing capture of incoming data. Mahanaxar addresses this problem. Mahanaxar guarantees streaming real-time data capture at (nearly) the full rate of the raw device, allows concurrent read and write access to the device on a best-effort basis without interrupting the data capture, and retains data as long as possible given the available storage. It has built in mechanisms for reliability and indexing, can scale to meet arbitrary bandwidth requirements, and handles both small and large data elements equally well. Results from our prototype implementation shows that Mahanaxar provides both better guarantees and better performance than traditional file systems.
As neural networks are more frequently used to solve problems in science and engineering, the methods used to incorporate scientific knowledge into these networks are becoming increasingly complex. Here, this work breaks down these complicated techniques into a set of basic strategies which can easily be applied to diverse situations. Several novel neural networks are built using the categories laid out in this work. These networks are tested on simulated data from a continuous stirred tank reactor (CSTR) model to evaluate the advantages provided by each network. The three points demonstrated in this work are: (1) architectural hybrid models can speed up convergence and reduce the amount of data necessary to train a model; (2) adding a physics-guided loss function can improve model generalization and make models more physically consistent; (3) using physics-guided initialization and transfer learning improves accuracy and speeds up convergence, but can harm generalizability if used incorrectly.
The role of big data in addressing the needs of the present healthcare system in US and rest of the world has been echoed by government, private, and academic sectors. There has been a growing emphasis to explore the promise of big data analytics in tapping the potential of the massive healthcare data emanating from private and government health insurance providers. While the domain implications of such collaboration are well known, this type of data has been explored to a limited extent in the data mining community. The objective of this paper is two fold: first, we introduce the emerging domain of big"healthcare claims data to the KDD community, and second, we describe the success and challenges that we encountered in analyzing this data using state of art analytics for massive data. Specically, we translate the problem of analyzing healthcare data into some of the most well-known analysis problems in the data mining community, social network analysis, text mining, and temporal analysis and higher order feature construction, and describe how advances within each of these areas can be leveraged to understand the domain of healthcare. Each case study illustrates a unique intersection of data mining and healthcare with a common objective of improving the cost-care ratio by mining for opportunities to improve healthcare operations and reducing hat seems to fall under fraud, waste,and abuse.
We are currently exploring data analysis methods for their ability to strengthen the synthesis and evaluation of information generated within domestic and international safeguard regimes. Safeguard data typically includes rich heterogenous datasets amenable to advanced data analytics methods, such as machine learning. We have converted transactional data containing both numerical and categorical attributes from a domestic nuclear material control and accountability (NMC&A) system into a low-dimensional numerical structure via linear principal component analysis. This data representation allows for global structure discovery via cluster analysis, which can characterize the typical behavior of each of the primary types of transaction events. Furthermore, the structure of the top principal components captures the “typical behavior” of the data and is thus amenable to anomaly detection through statistical hypothesis testing. We explored this capability by generating erroneous permutations of the data and computing the Q-residual quantity increase associated with the information loss when these data are projected into the low-dimensional principal component analysis space representative of typical transactions.Future work will focus on identifying data transformations (e.g., graph networks) that more closely align with the inherent structure of these transactions to explain more salient information and disambiguate the underlying nuclear process—in this case, irradiation and reprocessing—from artifacts of NMC&A system transactional record keeping data.
In the world of Internet of Things (IoT) networks, where devices are constantly communicating, keeping them secure from cyber threats is critical. This paper introduces a novel approach to detecting unusual and potentially harmful activities in these networks using graph neural networks (GNNs). We combine two specific types of GNNs-GraphSAGE and graph attention networks (GAT)-to create a model that understands and represents the behaviors and interactions in a network. GraphSAGE creates an embedding of network activities by examining local data interactions, while GAT directs the model's focus to the most critical interactions. By integrating these two methods in a single model that considers different types of interactions (both host and flow nodes), we aim to create a system that accurately represents the current state of a network and can also spot anomalies effectively while reducing false positives and negatives. Our innovative approach has demonstrated promising results, achieving an accuracy of 98% on the UNSW-NB15 dataset, significantly outperforming standalone GraphSAGE and GAT models. This underscores its potential as a robust framework for securing IoT networks against cyber threats and anomalies.
This dataset was used to assess the performance of a compression algorithm for Soil Deformation and Temperature Monitoring data. The development of this technology is presented in the article named time division duplex (TDD) "TDD LoRa and Delta Encoding in Low-Power Networks of Environmental Sensor Arrays for Temperature and Deformation Monitoring" and published in Springer's Journal on Signal Processing Systems. [Time Division Duplex (TDD); Long Range (LoRa)]Field experiment data includes: 1) soil temperature measurements, 2) snow temperature data, 3) soil temperature and deformation data. In these files, the first column contains the time (UTC), followed by the sensor's battery voltage and temperature and acceleration values (X, Y, Z) in subsequent columns. The measurements were acquired every 15 minutes. Lists included for probe, datalogger, and packaging components. The two temperature data sets were acquired at a site at the East River Watershed, Colorado, while the temperature/deformation data were collected at the Teller Road (mile 27) site near Nome, Alaska. The package includes 8 comma-separated files (.csv) and one .pdf file.The Next-Generation Ecosystem Experiments: Arctic (NGEE Arctic), was a research effort to reduce uncertainty in Earth System Models by developing a predictive understanding of carbon-rich Arctic ecosystems and feedbacks to climate. NGEE Arctic was supported by the Department of Energy's Office of Biological and Environmental Research.The NGEE Arctic project had two field research sites: 1) located within the Arctic polygonal tundra coastal region on the Barrow Environmental Observatory (BEO) and the North Slope near Utqiagvik (Barrow), Alaska and 2) multiple areas on the discontinuous permafrost region of the Seward Peninsula north of Nome, Alaska.Through observations, experiments, and synthesis with existing datasets, NGEE Arctic provided an enhanced knowledge base for multi-scale modeling and contributed to improved process representation at global pan-Arctic scales within the Department of Energy's Earth system Model (the Energy Exascale Earth System Model, or E3SM), and specifically within the E3SM Land Model component (ELM).
Enterprise networks of scale are complex, dynamic computing environments that respond to evolv- ing business objectives and requirements. Characteriz- ing system behaviors in these environments is essential for network management and cyber security operations. Characterization of system’s communication is typical and is supported using network flow information (NetFlow). Related work has characterized behavior using theoretical graph metrics; results are often difficult to interpret by enterprise staff. We propose a different approach, where flow information is mapped to sets of tags that contextualize the data in terms of network principals and enterprise concepts. Frequent patterns are then extracted and are expressed as behaviors. Behaviors can be com- pared, identifying systems expressing similar behaviors. We evaluate the approach using flow information collected by a third party.