Search NASA⌕ Search

SEARCH · Search NASA

Results for “deep generative models”

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 1,081 records · Page 60

MaLTESE: Large-Scale Simulation-Driven Machine Learning for Transient Driving Cycles

Optimal engine operation during a transient driving cycle is the key to achieving greater fuel economy, engine efficiency, and reduced emissions. In order to achieve continuously optimal engine operation, engine calibration methods use a combination of static correlations obtained from dynamometer tests for steady-state operating points and road and/or track performance data. As the parameter space of control variables, design variable constraints, and objective functions increases, the cost and duration for optimal calibration become prohibitively large. In order to reduce the number of dynamometer tests required for calibrating modern engines, a large-scale simulation-driven machine learning approach is presented in this work. A parallel, fast, robust, physics-based reduced-order engine simulator is used to obtain performance and emission characteristics of engines over a wide range of control parameters under various transient driving conditions (drive cycles). We scale the simulation up to 3,906 nodes of the Theta supercomputer at the Argonne Leadership Computing Facility to generate data required to train a machine learning model. The trained model is then used to predict various engine parameters of interest, and the results are compared with those predicted by the engine simulator. Our results show that a deepneural-network-based surrogate model achieves high accuracy: Pearson product-moment correlation values larger than 0.99 and mean absolute percentage error within 1.07% for various engine parameters such as exhaust temperature, exhaust pressure, nitric oxide, and engine torque. Once trained, the deep-neural-network-based surrogate model is fast for inference: it requires about 16 mu s for predicting the engine performance and emissions for a single design configuration compared with about 0.5 s per configuration with the engine simulator. Moreover, we demonstrate that transfer learning and retraining can be leveraged to incrementally retrain the surrogate model to cope with new configurations that fall outside the training data space.

deep learning↗

Kinetics of D/H isotope fractionation between molecular hydrogen and water

At equilibrium, the D/H isotope fractionation factor between H 2 and H 2 O (α H 2 O-H 2 (eq) ) is a sensitive indicator of temperature, and has been used as a geothermometer for natural springs and gas discharges. However, δD H 2 measured in spring waters may underestimate subsurface temperatures of origin due to partial isotopic re-equilibration during ascent and cooling. In this work, we present new experimental data on the kinetics of D–H exchange for H 2 dissolved in liquid water at temperatures below 100 °C. Comparing these results with published exchange rates obtained from gas phase experiments (100–400 °C), we derive a consistent activation energy of 52 kJ/mol, and the following rate expressions; ln k =9.186 - 6298/ T and k 1 = 9764.61[H 2 O] e-6298/ T where T is absolute temperature (K), k is the universal rate constant ([L/mol]/hr), and k 1 is a pseudo-first-order constant (hr -1 ) applicable to water-dominated terrestrial systems by constraining [H 2 O] as the density of H 2 O (in mol/L) at the P-T of interest. The density-dependent rate constant accounts for the kinetic disparity of D–H exchange with H 2 when dissolved in liquid H 2 O relative to a gas/steam phase, exemplifed by 1/ k 1 at 100 °C of ~2 days in liquid, versus ~7 yrs in saturated steam. This difference may explain the high variability of δDH 2 observed in fumarolic gases. Fluids convecting in the crust frequently reach T > 225 °C, where isotopic equilibrium is rapidly attained (<1 hr). We compare fractionation factors measured in natural fluids (α OBS ) with values expected for equilibrium at the T of acquisition. Where these values differ, we use kinetic models to estimate cooling rates during upward advection that account for the observed disequilibrium. Models fit to fluids from Yellowstone Park and the Lost City (deep-sea) vent field, both recovered at ~90 °C, require respective transit times of ~7 hrs and ~11 days between higher temperature reaction zones and the surface. Using estimates of subsurface depths of origin, however, suggests similar mean fluid flow rates (10 s of meters/hr). Additional complications must be considered when interpreting the δDH 2 of lower-temperature effluent. When applied to data from deep-sea hydrothermal systems, our kinetic models indicate microbial catalysis accelerates D–H exchange once fluids cool below ~60 °C. The H 2 measured in both continental alkaline springs and fracture fluids from Precambrian shield rock is likely produced at T < 100 °C, through processes such as serpentinization. In these settings, δDH 2 values appear closer to equilibrium with H 2 O than those from geothermal systems. Considering kinetic isotope effects may yield H 2 that is out of equilibrium when generated at lower temperatures, we calculate maximum (isothermal) times to apparent isotopic equilibrium of 1.3 yrs at 50 °C, 9 yrs at 25 °C, and 35 yrs at 10 °C. A similar calculation applied to Antarctic brines (-13 °C), where measured δDH 2 is far from equilibrium, yields ~350 yrs. This time is shorter than the fluids have been isolated (2.8 ka), suggesting kinetic isotope effects associated with H 2 destruction or loss via diffusion may also be possible.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Future energy infrastructure, energy platform and energy storage

Today fossil energy dominates energy consumption across the world. There has been an increasing momentum to reduce fossil energy consumption and increase renewable energy utilization to more than 70%. Such high penetrations of distributed renewable resources bring large uncertainty and complexity that cannot be easily handled by the current infrastructure. For the last two decades, telecommunication industry has gone through a similar transition from centralized to distributed systems, and more recently to platform approaches, which become the corner stone of today’s vibrant platform economy. Based on telecommunication history, we believe that a platform-based approach, called the energy platform, is a viable solution for addressing the renewable energy challenges. The energy platform consists of an array of computational algorithms, sensing and control technologies for key industry, energy generators and users to jointly manage and control the complex energy infrastructure. It includes the following key components: (1) the hardware and software to generate, store, control and transmit electricity/data (the energy cloud), (2) the digital platforms and tools for key energy producers, digital, service and communication industry to share the assets, participate in and manage the complex infrastructure, and for customers to participate in energy transactions, (3) the transaction platform for new service and business models. Additionally, the energy platform requires breakthroughs in large scale energy storage and many other areas including efficient power electronics, sensors and controls, new mathematical and computational tools, and deep integration of energy technologies and information sciences to control and stabilize such complex chaotic systems.

25 ENERGY STORAGE↗

Evolution of Precipitation Particle Size Distributions within MC3E Systems and its Impact on Aerosol-Cloud-Precipitation Interactions: Final Report

This is a multi-institutional, collaborative project using observations and modeling to study the evolution (e.g. formation and growth) of hydrometeors in continental convective clouds. Our contribution was in data analysis for the generation of high-value cloud and precipitation products and derive cloud statistics for model validation. There are two areas in data analysis that we contributed: i) the development of novel, state-of-the-art dual-wavelength radar algorithms for the retrieval of cloud microphysical properties and ii) the evaluation of large domain, high-resolution models using comprehensive multi-sensor observations. Our research group developed statistical summaries from numerous sensors and developed retrievals of vertical air motion in deep convection.

54 ENVIRONMENTAL SCIENCES↗

Particle modelling of magnetically confined oxygen plasma in low pressure radio frequency discharge

The main objective of this paper is the modelling and simulation of a radio frequency (RF) discharge in oxygen at low pressure and at room temperature, including the effect of crossed electric and magnetic fields for generation and confinement of oxygen plasma. The particle model takes into account one axial dimension along the electric field axis and three velocity components during the Monte Carlo treatment of the collisions between charged particles and background gas. The simulation by this developed code allows us not only to determine the electrodynamics characteristics of the RF discharge, but also to obtain kinetics and energetic description of reactive oxygen plasma at low pressure. These information are very important for the control of the deep reactive-ion etching technology of the silicon to manufacture capacitors with high density and for the deposition thick insulating films or thick metal to manufacture micro-coils. The simulation conditions are as follows: RF peak voltage of 200 V, frequency of 13.56 MHz, crossed magnetic field varying from 0 to 50 Gauss, and oxygen pressure of 13.8 Pa. In the presence of magnetic field, the results show an increase of the plasma density, a decrease of the electron mean energy, and also a reduction of the ratio between electron density and positive ion density. Finally in order to validate, the results are successfully compared with measurements already carried out in the literature. The conditions of comparison are from 100 to 300 V of the peak voltage at 13.56 MHz under a pressure of 13.8 Pa and a gap distance of 2.5 cm.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

BUOYANT MAGNETIC LOOPS IN A GLOBAL DYNAMO SIMULATION OF A YOUNG SUN

The current dynamo paradigm for the Sun and Sun-like stars places the generation site for strong toroidal magnetic structures deep in the solar interior. Sunspots and starspots on Sun-like stars are believed to arise when sections of these magnetic structures become buoyantly unstable and rise from the deep interior to the photosphere. Here, we present the first three-dimensional global magnetohydrodynamic (MHD) simulation in which turbulent convection, stratification, and rotation combine to yield a dynamo that self-consistently generates buoyant magnetic loops. We simulate stellar convection and dynamo action in a spherical shell with solar stratification, but rotating three times faster than the current solar rate. Strong wreaths of toroidal magnetic field are realized by dynamo action in the convection zone. By turning to a dynamic Smagorinsky model for subgrid-scale turbulence, we here attain considerably reduced diffusion in our simulation. This permits the regions of strongest magnetic field in these wreaths to rise toward the top of the convection zone via a combination of magnetic buoyancy instabilities and advection by convective giant cells. Such a global simulation yielding buoyant loops represents a significant step forward in combining numerical models of dynamo action and flux emergence.

79 ASTRONOMY AND ASTROPHYSICS↗

OpenUniverse2024: a shared, simulated view of the sky for the next generation of cosmological surveys

The OpenUniverse2024 simulation suite is a cross-collaboration effort to produce matched simulated imaging for multiple surveys as they would observe a common simulated sky. Both the simulated data and associated tools used to produce it are intended to uniquely enable a wide range of studies to maximize the science potential of the next generation of cosmological surveys. We have produced simulated imaging for approximately 70 deg 2 of the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST) Wide-Fast-Deep survey and the Nancy Grace Roman Space Telescope High-Latitude Wide-Area Survey, as well as overlapping versions of the ELAIS-S1 Deep-Drilling Field for LSST and the High-Latitude Time-Domain Survey for Roman. OpenUniverse2024 includes (i) an early version of the updated extragalactic model called Diffsky, which substantially improves the realism of optical and infrared photometry of objects, compared to previous versions of these models; (ii) updated transient models that extend through the wavelength range probed by Roman and Rubin; and (iii) improved survey, telescope, and instrument realism based on up-to-date survey plans and known properties of the instruments. It is built on a new and updated suite of simulation tools that improves the ease of consistently simulating multiple observatories viewing the same sky. The approximately 400 TB of synthetic survey imaging and simulated universe catalogs are publicly available, and we preview some scientific uses of the simulations.

large-scale structure of Universe↗

Sea quark transverse momentum distributions and dynamical chiral symmetry breaking

Recent theoretical studies have provided new insight into the intrinsic transverse momentum distributions of valence and sea quarks in the nucleon at a low scale. The valence quark transverse momentum distributions (q - qbar) are governed by the nucleon's inverse hadronic size R{sup -1} ~ 0.2 GeV and drop steeply at large p{sub T}. The sea quark distributions (qbar) are in large part generated by non-perturbative chiral-symmetry breaking interactions and extend up to the scale rho{sup -1} ~ 0.6 GeV. These findings have many implications for modeling the initial conditions of perturbative QCD evolution of TMD distributions (starting scale, shape of p{sub T}. distributions, coordinate-space correlation functions). The qualitative difference between valence and sea quark intrinsic p{sub T}. distributions could be observed experimentally, by comparing the transverse momentum distributions of selected hadrons in semi-inclusive deep-inelastic scattering, or those of dileptons produced in pp and pbar-p scattering.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Overview of Glenn Mechanical Components Branch Research

Mr. James Zakrajsek, chief of the Mechanical Components Branch, gave an overview of research conducted by the branch. Branch members perform basic research on mechanical components and systems, including gears and bearings, turbine seals, structural and thermal barrier seals, and space mechanisms. The research is focused on propulsion systems for present and advanced aerospace vehicles. For rotorcraft and conventional aircraft, we conduct research to develop technology needed to enable the design of low noise, ultra safe geared drive systems. We develop and validate analytical models for gear crack propagation, gear dynamics and noise, gear diagnostics, bearing dynamics, and thermal analyses of gear systems using experimental data from various component test rigs. In seal research we develop and test advanced turbine seal concepts to increase efficiency and durability of turbine engines. We perform experimental and analytical research to develop advanced thermal barrier seals and structural seals for current and next generation space vehicles. Our space mechanisms research involves fundamental investigation of lubricants, materials, components and mechanisms for deep space and planetary environments.

Zakrajsek, James↗

Agent-based modeling of microbes in space

Space is tough on organisms. Microorganisms traveling to space experience stress from environmental features such as ionizing radiation and lack of normal microgravity; however, much remains unknown about the mechanisms by which those environmental features affect microbial physiology. Microbes experience changes in gravity not directly but rather through changes in their fluid environment, and deep-space particle radiation causes cell damage that is complex but rare. Computational modeling at the single-cell level (agent-based modeling) can allow us to probe the spatially heterogeneous processes that characterize space stresses, to gain insight into the relationships of microbial cells with their environments and with each other. Here we present two software packages for simulating microbial population dynamics in space conditions: CAMDLES and AMMPER. Microbes growing in liquid culture medium in the microgravity of an orbital space station experience a quiescent, poorly-mixed fluid environment. CAMDLES (CFD-DEM Artificial Microgravity Developments for Living Ecosystem Simulation) simultaneously simulates biological, chemical, and mechanical processes to predict microbial ecological dynamics in microgravity, and in the rotating culture vessels used to create an artificial microgravity environment in the lab. Initial results demonstrate that the growth of a cross-feeding microbial consortium, dependent on the exchange of soluble metabolites, is sensitive to the initial spatial distribution of cells, and grows differently in real versus artificial microgravity. Microbial populations exposed to deep-space radiation experience spatially and temporally heterogeneous damage from the traversal of high-energy particles. AMMPER (Agent-Based Model for Microbial Populations Exposed to Radiation) pairs a 3d model of energy deposition along a radiation particle track with a microbial population growth and damage model to predict the effects of localized radiation damage on population-level responses. It includes a user-friendly graphical interface. AMMPER results agree with experimental data indicating that indirect effects of radiation (reactive oxygen species generation, metabolic impairment) have a greater impact on microorganisms than direct effects (DNA damage).

Jessica A Lee↗

Multiscale Modeling of Reconstructed Tricalcium Silicate using NASA Multiscale Analysis Tool

To study microstructure characteristics of cementitious materials hydrated in space; previously, cement binder formations were processed under microgravity conditions and was further compared against ground-based experiments. For accurate estimation of process-structure-property linkage, particularly on samples hydrated in the microgravity environment, it is desired to have a high-fidelity volumetric representation of the microstructure. However, owing to small sample size and high porosity of the space-returned samples, conventional experimental characterization techniques are not viable. Hence, a deep learning-based reconstruction algorithm was employed to obtain high fidelity 3D volumes from sparse high resolution 2D Scanning Electron Microscopy (SEM) images, as inputs to micromechanics-based modeling. This machine learning-based reconstruction methodology validated against low-order statistical descriptors, captured the microstructural topology of both sample types (ground, 1g and microgravity, μg). Due to the lack of gravity, hydration products of the samples processed in space differed from those processed-on ground. Such AI-generated virtual samples were analyzed in a multiscale recursive micromechanics approach using the NASA Multiscale Analysis Tool (NASMAT). Here, we present a methodology to rapidly integrate and evaluate these AI-generated volumes in NASMAT. The synthesized microstructural volumes are directly employed as Representative Volume Elements (RVEs) to preserve the fidelity (1 pixel = 0.54 m). Invariably, analysis of such largescale problems (5123 voxels) requires huge amount of computational resources. By taking advantage of the NASMAT architecture, we also focused on systematic multiscale integration of these AI-reconstructed virtual volumes to reduce the computational demands. In this work, this methodology is demonstrated on the ground-based, 1g samples. The estimated stiffness value of 15.90 GPa is comparable to experimentally obtained modulus of hydrated tricalcium silicate sample. The workflow presented here paves the way for utilizing the NASMAT tool to perform multiscale analyses of other multi-phase material systems using either 3D virtual datasets synthesized using AI or obtained via micro-CT.

Machine Learning↗

Evaluation of SNS Beamline Shielding Configurations using MCNPX Accelerated by ADVANTG

Shielding analyses for the Spallation Neutron Source (SNS) at Oak Ridge National Laboratory pose significant computational challenges, including highly anisotropic high-energy sources, a combination of deep penetration shielding and an unshielded beamline, and a desire to obtain well-converged nearly global solutions for mapping of predicted radiation fields. The majority of these analyses have been performed using MCNPX with manually generated variance reduction parameters (source biasing and cell-based splitting and Russian roulette) that were largely based on the analyst's insight into the problem specifics. Development of the variance reduction parameters required extensive analyst time, and was often tailored to specific portions of the model phase space. We previously applied a developmental version of the ADVANTG code to an SNS beamline study to perform a hybrid deterministic/Monte Carlo analysis and showed that we could obtain nearly global Monte Carlo solutions with essentially uniform relative errors for mesh tallies that cover extensive portions of the model with typical voxel spacing of a few centimeters. The use of weight window maps and consistent biased sources produced using the FW-CADIS methodology in ADVANTG allowed us to obtain these solutions using substantially less computer time than the previous cell-based splitting approach. While those results were promising, the process of using the developmental version of ADVANTG was somewhat laborious, requiring user-developed Python scripts to drive much of the analysis sequence. In addition, limitations imposed by the size of weight-window files in MCNPX necessitated the use of relatively coarse spatial and energy discretization for the deterministic Denovo calculations that we used to generate the variance reduction parameters. We recently applied the production version of ADVANTG to this beamline analysis, which substantially streamlined the analysis process. We also tested importance function collapsing (in space and energy) capabilities in ADVANTG. These changes, along with the support for parallel Denovo calculations using the current version of ADVANTG, give us the capability to improve the fidelity of the deterministic portion of the hybrid analysis sequence, obtain improved weight-window maps, and reduce both the analyst and computational time required for the analysis process.

43 PARTICLE ACCELERATORS↗

Real-time sparse-sampled Ptychographic imaging through deep neural networks

Ptychography has rapidly grown in the fields of X-ray and electron imaging for its unprecedented ability to achieve nano or atomic scale resolution while simultaneously retrieving chemical or magnetic information from a sample. A ptychographic reconstruction is achieved by means of solving a complex inverse problem that imposes constraints both on the acquisition and on the analysis of the data, which typically precludes real-time imaging due to computational cost involved in solving this inverse problem. In this work we propose PtychoNN, a novel approach to solve the ptychography reconstruction problem based on deep convolutional neural networks. We demonstrate how the proposed method can be used to predict real-space structure and phase at each scan point solely from the corresponding far-field diffraction data. The presented results demonstrate how PtychoNN can effectively be used on experimental data, being able to generate high quality reconstructions of a sample up to hundreds of times faster than state-of-the-art ptychography reconstruction solutions once trained. By surpassing the typical constraints of iterative model-based methods, we can significantly relax the data acquisition sampling conditions and produce equally satisfactory reconstructions. Besides drastically accelerating acquisition and analysis, this capability can enable new imaging scenarios that were not possible before, in cases of dose sensitive, dynamic and extremely voluminous samples.

Cherukara, Mathew J↗

Automatic Detection and Classification of Radio Galaxy Images by Deep Learning

Abstract Surveys conducted by radio astronomy observatories, such as SKA, MeerKAT, Very Large Array, and ASKAP, have generated massive astronomical images containing radio galaxies (RGs). This generation of massive RG images has imposed strict requirements on the detection and classification of RGs and makes manual classification and detection increasingly difficult, even impossible. Rapid classification and detection of images of different types of RGs help astronomers make full use of the observed astronomical image data for further processing and analysis. The classification of FRI and FRII is relatively easy, and there are more studies and literature on them at present, but FR0 and FRI are similar, so it is difficult to distinguish them. It poses a greater challenge to image processing. At present, deep learning has made breakthrough progress in the field of image analysis and processing and has preliminary applications in astronomical data processing. Compared with classification algorithms that can only classify galaxies, object detection algorithms that can locate and classify RGs simultaneously are preferred. In target detection algorithms, YOLOv5 has outstanding advantages in the classification and positioning of small targets. Therefore, we propose a deep-learning method based on an improved YOLOv5 object detection model that makes full use of multisource data, combining FIRST radio with SDSS optical image data, and realizes the automatic detection of FR0, FRI, and FRII RGs. The innovation of our work is that on the basis of the original YOLOv5 object detection model, we introduce the SE Net attention mechanism, increase the number of preset anchors, adjust the network structure of the feature pyramid, and modify the network structure, thereby allowing our model to demonstrate galaxy classification and position detection effects. Our improved model produces satisfactory results, as evidenced by experiments. Overall, the mean average precision (mAP@0.5) of our improved model on the test set reaches 89.4%, which can determine the position (R.A. and decl.) and automatically detect and classify FR0s, FRIs, and FRIIs. Our work contributes to astronomy because it allows astronomers to locate FR0, FRI, and FRII galaxies in a relatively short time and can be further combined with other astronomically generated data to study the properties of these galaxies. The target detection model can also help astronomers find FR0s, FRIs, and FRIIs in future surveys and build a large-scale star RG catalog. Moreover, our work is also useful for the detection of other types of galaxies.

Astronomy & Astrophysics↗

Experimental Soil Warming Impacts Soil Moisture and Plant Water Stress and Thereby Ecosystem Carbon Dynamics (Blodgett, CA)

This dataset contains data on daily soil temperature, moisture and flux, and soil carbon stock and root biomass across a soil profile down to 100 cm depth at Blodgett Forest Research Station, CA, USA. These data were generated to determine if modeling of an experimental soil warming of 4C showed increased soil CO2 emissions and changes in bulk soil carbon stocks with depth consistent with field observations, as part of the study: Riley et al. (2025) Experimental Soil Warming Impacts Soil Moisture and Plant Water Stress and Thereby Ecosystem Carbon Dynamics in Journal of Advances in Modeling Earth Systems. This research was performed within the framework of the TES Belowground Biogeochemistry SFA project, in particular association with a 1 m-deep experimental soil heating experiment at the University of California Blodgett Forest Research Station, California (120 ° 39′40′′W; 38 ° 54′43′′N). Continuous data were collected at the plot level, and bulk soil carbon and root biomass were sampled once a year from each plot from 0-100 cm, in 10 cm intervals. Measurements relevant to the current study include soil temperature and soil volumetric water content measured continuously at multiple depths in the top meter; fine root biomass and SOC stocks measured from annual soil cores. Soil flux was continuously monitored using a LI-8100 Automated CO2 Flux System in conjunction with the LI-8150 Multiplexer (Licor, Nebraska, USA). Soil flux was determined using SoilFluxPro software, with flux values showing an R² fit of less than 0.9 being excluded from the analysis. Data were collected from each paired plot (1-3): one control (C) and one heated (H).

54 ENVIRONMENTAL SCIENCES↗

Deep Learning for Full Waveform Inversion of Elastic Active-Source Seismic Data to Estimate P-Wave Velocity Models

Seismic imaging methods are critical for Global Security and Energy & Homeland Security missions and activities that rely on subsurface characterization, but traditional methods remain computationally expensive and require significant labor hours and expertise to execute. Within the past few years, machine learning (ML), namely deep learning (DL), has been used to develop data-driven end-to-end full waveform inversion (FWI) methods to estimate 2D P-wave velocity (Vp) models in a fraction of the time as conventional FWI. These methods, however, are trained on simplistic acoustic wave seismic data and Vp models that are not realistic nor representative of real-world observations, leaving a large gap between the state-of-the-art and deployable, feasible, and practical DL FWI methods. Here, we generate a synthetic active-source, 3D, elastic wave seismic data set and a variety of Vp models with realistic geologic structure for training DL FWI methods. We evaluate six different methods that have performed well for acoustic DL FWI or medical imaging tasks using our more realistic dataset. We find that these six trained models do not match the performance of published acoustic end-to-end DL FWI methods, indicating more training data may be needed, physics may need to be incorporated to achieve good accuracy at the sacrifice of the end-to-end advantage, and/or novel methods need to be developed to enable end-to-end DL FWI methods to perform well for real-world seismic data.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Computer-aided Abnormality Detection in Chest Radiographs in a Clinical Setting via Domain-adaptation

Deep learning (DL) models are being deployed at medical centers to aid radiologists for diagnosis of lung conditions from chest radiographs. Such models are often trained on a large volume of publicly available labeled radiographs. These pre-trained DL models’ ability to generalize in clinical settings is poor because of the changes in data distributions between publicly available and privately held radiographs. In chest radiographs, the heterogeneity in distributions arises from the diverse conditions in X-ray equipment and their configurations used for generating the images. In the machine learning community, the challenges posed by the heterogeneity in the data generation source is known as domain shift, which is a mode shift in the generative model. In this work, we introduce a domain-shift detection and removal method to overcome this problem. Our experimental results show the proposed method’s effectiveness in deploying a pre-trained DL model for abnormality detection in chest radiographs in a clinical setting.

Dubey, Abhishek↗

Regional price targets appropriate for advanced coal extraction

A methodology is presented for predicting coal prices in regional markets for the target time frames 1985 and 2000 that could subsequently be used to guide the development of an advanced coal extraction system. The model constructed is a supply and demand model that focuses on underground mining since the advanced technology is expected to be developed for these reserves by the target years. Coal reserve data and the cost of operating a mine are used to obtain the minimum acceptable selling price that would induce the producer to bring the mine into production. Based on this information, market supply curves can be generated. Demand by region is calculated based on an EEA methodology that emphasizes demand by electric utilities and demand by industry. The demand and supply curves are then used to obtain the price targets. The results show a growth in the size of the markets for compliance and low sulphur coal regions. A significant rise in the real price of coal is not expected even by the year 2000. The model predicts heavy reliance on mines with thick seams, larger block size and deep overburden.

Terasawa, K. L.↗