Search NASA⌕ Search

SEARCH · Search NASA

Results for “deep direct use”

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 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Effects of head modeling errors on the spatial frequency representation of MEG

Abstract Objectives. We aim to investigate the effects of head model inaccuracies on signal and source reconstruction accuracies for various sensor array distances to the head. This allows for the assessment of the importance of head modeling for next-generation magnetoencephalography (MEG) sensors, optically-pumped magnetometers (OPM). Approach. A 1-shell boundary element method (BEM) spherical head model with 642 vertices of radius 9 cm and conductivity of 0.33 S m −1 was defined. The vertices were then randomly perturbed radially up to 2%, 4%, 6%, 8% and 10% of the radius. For each head perturbation case, the forward signal was calculated for dipolar sources located at 2 cm, 4 cm, 6 cm and 8 cm from the origin (center of the sphere), and for a 324 sensor array located at 10 cm to 15 cm from the origin. Equivalent current dipole (ECD) source localization was performed for each of these forward signals. The signal for each perturbed spherical head case was then analyzed in the spatial frequency domain, and the signal and ECD errors were quantified relative to the unperturbed case. Main results. In the noiseless and high signal-to-noise ratio (SNR) case of approximately ≥6 dB, inaccuracies in our spherical BEM head conductor models lead to increased signal and ECD inaccuracies when sensor arrays are placed closer to the head. This is true especially in the case of deep and superficial sources. In the noisy case however, the higher SNR for closer sensor arrays allows for an improved ECD fit and outweighs the effects of head geometry inaccuracies. Significance. OPMs may be placed directly on the head, as opposed to the more commonly used superconducting quantum interference device sensors which must be placed a few centimeters away from the head. OPMs thus allow for signals of higher spatial resolution to be captured, resulting in potentially more accurate source localizations. Our results suggest that an increased emphasis on accurate head modeling for OPMs may be necessary to fully realize its improved source localization potential.

Yeo, Wan-Jin (ORCID:0000000241417123)↗
DOE OSTI2025

Comparison of Coupled and Uncoupled Modeling of Floating Wind Farms with Shared Anchors

As design options for floating wind farms continue to be explored, shared (or multiline) anchors that secure mooring lines from multiple turbines remain a promising technology that can potentially reduce the number of anchors and overall mooring costs. This study evaluates two methods for analyzing the loads on shared anchors: one in which floating offshore wind turbines are simulated individually (using the software OpenFAST), and one in which an entire floating wind farm is simulated collectively (using the software FAST.Farm). A three-line shared anchor is evaluated for multiple loading scenarios in deep water, using the International Energy Agency 15 MW turbine on the VolturnUS-S semisubmersible platform. While the two methods produce broadly comparable results, the coupled wave loading on platforms within the farm results in wave force cancellations and amplifications that decrease multiline force directional ranges and increase multiline force extreme values (up to 7%) and standard deviations (up to 11%) for wave-driven load cases. The inclusion of wakes in FAST.Farm also reduces the net load on the shared anchor due to the velocity deficit, leading to larger differences between OpenFAST and FAST.Farm (up to 3% difference in mean loads) for load cases with operational turbines.

17 WIND ENERGY↗
DOE OSTI2021

Determining fast-S and slow-S propagation directions with SV-P data produced by buried explosives and recorded with vertical geophones

We have evaluated the concept of practicing S-wave reflection seismology with legacy 3D seismic data generated by a P-wave source and recorded with only vertical geophones. This type of S-wave imaging is based on the principle that seismic P-wave sources not only produce a downgoing illuminating P wavefield, but they also simultaneously produce a downgoing illuminating SV wavefield that, in almost all cases, is suitable for S-wave reflection imaging. The S-mode used in this study is the SV-P, or converted-P, mode. This mode involves a downgoing illuminating SV wavefield and an upgoing reflected P-mode that is recorded by vertical geophones. In flat-layered stratigraphy, the lengths of the SV and P raypaths in SV-P imaging are identical to the lengths of the SV and P raypaths in P-SV imaging with P-sources and 3C geophones. P-SV imaging of deep rocks has been practiced for more than two decades; SV-P imaging is a new concept. SV-P data should provide the same options for investigating deep rocks as do P-SV data. We have determined one of the equivalences between SV-P data extracted from vertical-geophone data and P-SV data extracted from horizontal geophones: that both modes react to azimuth-dependent variations in the S velocity in anisotropic rocks. Azimuthal variations in the SV-P traveltime can be used to define the polarization direction of the fast-S-wave mode, which is also the azimuth of the maximum horizontal stress (SHmax). Our investigation demonstrates a noninvasive method for monitoring changes in the SHmax azimuth across a CO 2 storage reservoir, or any targeted porous rock, as fluids are cycled into, and then out of, that rock’s pore space.

Geochemistry & Geophysics↗
DOE OSTI2022

Transfer learning for smart buildings: A critical review of algorithms, applications, and future perspectives

Smart buildings play a crucial role toward decarbonizing society, as globally buildings emit about one-third of greenhouse gases. In the last few years, machine learning has achieved a notable momentum that, if properly harnessed, may unleash its potential for advanced analytics and control of smart buildings, enabling the technique to scale up for supporting the decarbonization of the building sector. In this perspective, transfer learning aims to improve the performance of a target learner exploiting knowledge in related environments. The present work provides a comprehensive overview of transfer learning applications in smart buildings, classifying and analyzing 77 papers according to their applications, algorithms, and adopted metrics. The study identified four main application areas of transfer learning: (1) building load prediction, (2) occupancy detection and activity recognition, (3) building dynamics modeling, and (4) energy systems control. Furthermore, the review highlighted the role of deep learning in transfer learning applications that has been used in more than half of the analyzed studies. The paper also discusses how to integrate transfer learning in a smart building's ecosystem, identifying, for each application area, the research gaps and guidelines for future research directions.

Pinto, G↗
DOE OSTI2024

Graph Analytics for CEBAF Operations

We report on the progress achieved during a 2-year Laboratory Directed Research and Development (LDRD) project titled “Graph Analytics for CEBAF Operations”. The objective of this project is to leverage deep learning on graph representations of CEBAF’s injector beamline in order to create a tool for improving the efficiency of beam tuning tasks. Specifically, we use graphs to represent the injector beamline at any arbitrary date and time and invoke a graph neural network (GNN) to extract a low-dimensional, informative representation that can be visualized in two-dimensions. By analyzing years of operational data from the CEBAF archiver, good and bad regions of parameter space can be identified. The goal is to exercise this framework as a real-time tool to aid beam tuning, which represents the dominant source of machine downtime.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

Improving North American Wildfire Prediction by Integrating a Machine-Learning Fire Model in a Land Surface Model

Wildfires have shown increasing trends in both frequency and severity across the Contiguous United States (CONUS). However, process-based fire models have difficulties in accurately simulating the burned area over the CONUS due to a simplification of the physical process and cannot capture the interplay among fire, ignition, climate, and human activities. The deficiency of burned area simulation deteriorates the description of fire impact on energy balance, water budget, and carbon fluxes in the Earth System Models (ESMs). Alternatively, machine learning (ML) based fire models, which capture statistical relationships between the burned area and environmental factors, have shown promising burned area predictions and corresponding fire impact simulation. We develop a hybrid framework (ML4Fire-XGB) that integrates a pretrained eXtreme Gradient Boosting (XGBoost) wildfire model with the Energy Exascale Earth System Model (E3SM) land model (ELM) version 2.1. A Fortran-C-Python deep learning bridge is adapted to support online communication between ELM and the ML fire model. Specifically, the burned area predicted by the ML-based wildfire model is directly passed to ELM to adjust the carbon pool and vegetation dynamics after disturbance, which are then used as predictors in the ML-based fire model in the next time step. Evaluated against the historical burned area from Global Fire Emissions Database 5 from 2001-2020, the ML4Fire-XGB model outperforms process-based fire models in terms of spatial distribution and seasonal variations. Sensitivity analysis confirms that the ML4Fire-XGB well captures the responses of the burned area to rising temperatures. The ML4Fire-XGB model has proved to be a new tool for studying vegetation-fire interactions, and more importantly, enables seamless exploration of climate-fire feedback, working as an active component in E3SM.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Simulated wildfire burned area over the CONUS during 2001-2020

Wildfires have shown increasing trends in both frequency and severity across the Contiguous United States (CONUS). However, process-based fire models have difficulties in accurately simulating the burned area over the CONUS due to a simplification of the physical process and cannot capture the interplay among fire, ignition, climate, and human activities. The deficiency of burned area simulation deteriorates the description of fire impact on energy balance, water budget, and carbon fluxes in the Earth System Models (ESMs). Alternatively, machine learning (ML) based fire models, which capture statistical relationships between the burned area and environmental factors, have shown promising burned area predictions and corresponding fire impact simulation. We develop a hybrid framework (ML4Fire-XGB) that integrates a pretrained eXtreme Gradient Boosting (XGBoost) wildfire model with the Energy Exascale Earth System Model (E3SM) land model (ELM). A Fortran-C-Python deep learning bridge is adapted to support online communication between ELM and the ML fire model. Specifically, the burned area predicted by the ML-based wildfire model is directly passed to ELM to adjust the carbon pool and vegetation dynamics after disturbance, which are then used as predictors in the ML-based fire model in the next time step. Evaluated against the historical burned area from Global Fire Emissions Database 5 from 2001-2020, the ML4Fire-XGB model outperforms process-based fire models in terms of spatial distribution and seasonal variations. Sensitivity analysis confirms that the ML4Fire-XGB well captures the responses of the burned area to rising temperatures. The ML4Fire-XGB model has proved to be a new tool for studying vegetation-fire interactions, and more importantly, enables seamless exploration of climate-fire feedback, working as an active component in E3SM.

Liu, Ye↗
DOE OSTI2025

ELM2.1-XGBfire1.0: improving wildfire prediction by integrating a machine learning fire model in a land surface model

Wildfires have shown increasing trends in both frequency and severity across the contiguous United States (CONUS). However, process-based fire models have difficulties in accurately simulating the burned area over the CONUS due to a simplification of the physical process and cannot capture the interplay among fire, ignition, climate, and human activities. The deficiency of burned area simulation deteriorates the description of fire impact on energy balance, water budget, and carbon fluxes in the Earth system models (ESMs). Alternatively, fire models based on machine learning (ML), which capture statistical relationships between the burned area and environmental factors, have shown promising burned area predictions and corresponding fire impact simulation. We develop a hybrid framework (ELM2.1-XGBFire1.0) that integrates an eXtreme Gradient Boosting (XGBoost) wildfire model with the Energy Exascale Earth System Model (E3SM) land model (ELM) version 2.1. A Fortran–C–Python deep learning bridge is adapted to support online communication between ELM and the ML fire model. Specifically, the burned area predicted by the ML-based wildfire model is directly passed to ELM to adjust the carbon pool and vegetation dynamics after disturbance, which are then used as predictors in the ML-based fire model in the next time step. Evaluated against the historical burned area from Global Fire Emissions Database 5 from 2001–2019, the ELM2.1-XGBFire1.0 outperforms process-based fire models in terms of spatial distribution and seasonal variations. The ELM2.1-XGBFire1.0 has proven to be a new tool for studying vegetation–fire interactions and, more importantly, enables seamless exploration of climate–fire feedback, working as an active component of E3SM.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Knowledge-informed deep learning for hydrological model calibration: an application to Coal Creek Watershed in Colorado

Abstract. Deep learning (DL)-assisted inverse mapping has shown promise in hydrological model calibration by directly estimating parameters from observations. However, the increasing computational demand for running the state-of-the-art hydrological model limits sufficient ensemble runs for its calibration. In this work, we present a novel knowledge-informed deep learning method that can efficiently conduct the calibration using a few hundred realizations. The method involves two steps. First, we determine decisive model parameters from a complete parameter set based on the mutual information (MI) between model responses and each parameter computed by a limited number of realizations (∼50). Second, we perform more ensemble runs (e.g., several hundred) to generate the training sets for the inverse mapping, which selects informative model responses for estimating each parameter using MI-based parameter sensitivity. We applied this new DL-based method to calibrate a process-based integrated hydrological model, the Advanced Terrestrial Simulator (ATS), at Coal Creek Watershed, CO. The calibration is performed against observed stream discharge (Q) and remotely sensed evapotranspiration (ET) from the water year 2017 to 2019. Preliminary MI analysis on 50 realizations resulted in a down-selection of 7 out of 14 ATS model parameters. Then, we performed a complete MI analysis on 396 realizations and constructed the inverse mapping from informative responses to each of the selected parameters using a deep neural network. Compared with calibration using observations covering all time steps, the new inverse mapping improves parameter estimations, thus enhancing the performance of ATS forward model runs. The Nash–Sutcliffe efficiency (NSE) of streamflow predictions increases from 0.53 to 0.8 when calibrating against Q alone. Using ET observations, on the other hand, does not show much improvement on the performance of ATS modeling mainly due to both the uncertainty of the remotely sensed product and the insufficient coverage of the model ET ensemble in capturing the observation. By using observed Q only, we further performed a multiyear analysis and show that Q is best simulated (NSE > 0.8) by including in the calibration the dry-year flow dynamics that show more sensitivity to subsurface characteristics than the other wet years. Moreover, when continuing the forward runs till the end of 2021, the calibrated models show similar simulation performances during this evaluation period as the calibration period, demonstrating the ability of the estimated parameters in capturing climate sensitivity. Our success highlights the importance of leveraging data-driven knowledge in DL-assisted hydrological model calibration.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Automated Grain Boundary (GB) Segmentation and Microstructural Analysis in 347H Stainless Steel Using Deep Learning and Multimodal Microscopy

Austenitic 347H stainless steel offers superior mechanical properties and corrosion resistance required for extreme operating conditions such as high temperature. The change in microstructure due to composition and process variations is expected to impact material properties. Identifying microstructural features such as grain boundaries thus becomes an important task in the process-microstructure-properties loop. Applying convolutional neural network (CNN)-based deep learning models is a powerful technique to detect features from material micrographs in an automated manner. In contrast to microstructural classification, supervised CNN models for segmentation tasks require pixel-wise annotation labels. However, manual labeling of the images for the segmentation task poses a major bottleneck for generating training data and labels in a reliable and reproducible way within a reasonable timeframe. Microstructural characterization especially needs to be expedited for faster material discovery by changing alloy compositions. Here, in this study, we attempt to overcome such limitations by utilizing multimodal microscopy to generate labels directly instead of manual labeling. We combine scanning electron microscopy images of 347H stainless steel as training data and electron backscatter diffraction micrographs as pixel-wise labels for grain boundary detection as a semantic segmentation task. The viability of our method is evaluated by considering a set of deep CNN architectures. We demonstrate that despite producing instrumentation drift during data collection between two modes of microscopy, this method performs comparably to similar segmentation tasks that used manual labeling. Additionally, we find that naïve pixel-wise segmentation results in small gaps and missing boundaries in the predicted grain boundary map. By incorporating topological information during model training, the connectivity of the grain boundary network and segmentation performance is improved. Finally, our approach is validated by accurate computation on downstream tasks of predicting the underlying grain morphology distributions which are the ultimate quantities of interest for microstructural characterization.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Domain Aware Deep-learning Algorithms Integrated with Scientific-computing Technologies (DADAIST)

This technical report summarized the contribution of the DADAIST project funded by the Data Model Convergence Initiative via the Laboratory Directed Research and Development (LDRD) investments at Pacific Northwest National Laboratory (PNNL). Specifically, we report the development of the NeuroMANCER (Neural Modules with Adaptive Nonlinear Constraints and Efficient Regularizations), a new open-source Scientific Machine Learning library for formulating and solving parametric constrained optimization problems, physics-informed system identification, and parametric optimal control problems. NeuroMANCER is using differentiable programming to combine modern data-driven models and optimization modeling language into a coherent algorithmic and software framework. NeuroMANCER is a Pytorch-based framework and adopts much of its philosophy focused on research and development, rapid prototyping, and streamlined deployment. Strong emphasis is given to extensibility, interoperability with the PyTorch ecosystem, and quick adaptability to custom domain problems. Neuromancer repository contains a comprehensive library of differentiable modules, including custom activation functions, matrix factorizations, deep learning architectures, neural differential equations, differential equation solvers, implicit layers such as iterative solvers, high-level API for symbolic expressions, API for modeling and control of dynamical systems, and extensive set of tutorial code examples in the form of python scripts and jupyter notebooks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

AI-NERD: Elucidation of relaxation dynamics beyond equilibrium through AI-informed X-ray photon correlation spectroscopy

Abstract Understanding and interpreting dynamics of functional materials in situ is a grand challenge in physics and materials science due to the difficulty of experimentally probing materials at varied length and time scales. X-ray photon correlation spectroscopy (XPCS) is uniquely well-suited for characterizing materials dynamics over wide-ranging time scales. However, spatial and temporal heterogeneity in material behavior can make interpretation of experimental XPCS data difficult. In this work, we have developed an unsupervised deep learning (DL) framework for automated classification of relaxation dynamics from experimental data without requiring any prior physical knowledge of the system. We demonstrate how this method can be used to accelerate exploration of large datasets to identify samples of interest, and we apply this approach to directly correlate microscopic dynamics with macroscopic properties of a model system. Importantly, this DL framework is material and process agnostic, marking a concrete step towards autonomous materials discovery.

36 MATERIALS SCIENCE↗
DOE OSTI2024

EGS Reservoir Modeling for Developing Geothermal District Heating at Cornell University: Preprint

Cornell University is pursuing development of an enhanced geothermal system (EGS) for providing heating to its main campus in Upstate New York. A ~10,000 ft (~3 km) deep vertical observation well ("CUBO") was drilled in 2022 to characterize the subsurface using wellbore logging, borehole imaging, fluid sampling, mini-frac tests, coring and drill cutting analysis. Down-hole temperatures measured at 3 km depth are about 80 degrees C, sufficiently high for direct-use heating. The well drilled through generally low porosity and low permeability Paleozoic sedimentary formations and into metamorphic basement rock, encountered at about 9,400 ft depth. Leveraging subsurface data obtained through CUBO, we investigated technical feasibility and design requirements of a doublet well system with horizontal laterals connected to a fracture network created through hydraulic fracturing. The EGS reservoir is sized to provide a nominal heat output in the range 5 to 10 MWth of continuous heating over a 15 year-lifetime with limited thermal drawdown. We applied the Gringarten Multiple parallel fracture model, the Cornell Discrete Fracture Simulator FOXFEM and the commercial simulator ResFrac to estimate required heat transfer area and design a potential hydraulic stimulation treatment. Reservoir simulations indicate that, depending on fluid flow rate and injection temperature, 2 to 3 km2 of effective fracture heat transfer area is required to supply the target heat output of 5 to 10 MWth over 15 years.

district heating↗
DOE OSTI2025

EGS Reservoir Modeling for Developing Geothermal District Heating at Cornell University

Cornell University is pursuing development of an enhanced geothermal system (EGS) for providing heating to its main campus in Upstate New York. A ~10,000 ft (~3 km) deep vertical observation well ('CUBO') was drilled in 2022 to characterize the subsurface using wellbore logging, borehole imaging, fluid sampling, mini-frac tests, coring and drill cutting analysis. Down-hole temperatures measured at 3 km depth are about 80 degrees C, sufficiently high for direct-use heating. The well drilled through generally low porosity and low permeability Paleozoic sedimentary formations and into metamorphic basement rock, encountered at about 9,400 ft depth. Leveraging subsurface data obtained through CUBO, we investigated technical feasibility and design requirements of a doublet well system with horizontal laterals connected to a fracture network created through hydraulic fracturing. The EGS reservoir is sized to provide a nominal heat output in the range 5 to 10 MWth of continuous heating over a 15 year-lifetime with limited thermal drawdown. We applied the Gringarten Multiple parallel fracture model, the Cornell Discrete Fracture Simulator FOXFEM and the commercial simulator ResFrac to estimate required heat transfer area and design a potential hydraulic stimulation treatment. Reservoir simulations indicate that, depending on fluid flow rate and injection temperature, 2 to 3 km2 of effective fracture heat transfer area is required to supply the target heat output of 5 to 10 MWth over 15 years.

district heating↗
DOE OSTI2023

Cost and Life Cycle Analysis for Deep CO 2 Emissions Reduction for Steel Making: Direct Reduced Iron Technologies

Among heavy industrial sectors worldwide, the steel industry ranks first in carbon dioxide (CO 2 ) emissions. Technologies that produce direct reduced iron (DRI) enable the industry to reduce emissions or even approach net‐zero CO 2 emissions for steel production. Herein, comprehensive cradle‐to‐gate (CTG) life cycle analysis (LCA) and techno‐economic analysis (TEA) are used to evaluate the CO 2 emissions of three DRI technologies. Compared to the baseline of blast furnace and basic oxygen furnace (BF–BOF) technology for steel making, using natural gas (NG) to produce DRI has the potential to reduce CTG CO 2 emissions by 33%. When 83% or 100% renewable H 2 is used for DRI production, DRI technologies can potentially reduce CO 2 emissions by 57% and 67%, respectively, compared to baseline BF–BOF technology. However, the renewable H 2 application for DRI increases the levelized cost of steel (LCOS). When renewable natural gas (RNG) and clean electricity are used for steel production, the CTG CO 2 emissions of all the DRI technologies can potentially be reduced by more than 90% compared to the baseline BF–BOF technology, although the LCOS depends largely on the cost of RNG and clean electricity.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2020

GROWTH on S190814bv: Deep Synoptic Limits on the Optical/Near-infrared Counterpart to a Neutron Star–Black Hole Merger

On 2019 August 14, the Advanced LIGO and Virgo interferometers detected the high-significance gravitational wave (GW) signal S190814bv. The GW data indicated that the event resulted from a neutron star–black hole (NSBH) merger, or potentially a low-mass binary BH merger. Due to the low false-alarm rate and the precise localization (23 deg 2 at 90%), S190814bv presented the community with the best opportunity yet to directly observe an optical/near-infrared counterpart to an NSBH merger. To search for potential counterparts, the GROWTH Collaboration performed real-time image subtraction on six nights of public Dark Energy Camera images acquired in the 3 weeks following the merger, covering >98% of the localization probability. Using a worldwide network of follow-up facilities, we systematically undertook spectroscopy and imaging of optical counterpart candidates. Combining these data with a photometric redshift catalog, we ruled out each candidate as the counterpart to S190814bv and placed deep, uniform limits on the optical emission associated with S190814bv. For the nearest consistent GW distance, radiative transfer simulations of NSBH mergers constrain the ejecta mass of S190814bv to be M ej < 0.04 M ⊙ at polar viewing angles, or M ej < 0.03 M ⊙ if the opacity is κ < 2 cm 2 g -1 . Assuming a tidal deformability for the NS at the high end of the range compatible with GW170817 results, our limits would constrain the BH spin component aligned with the orbital momentum to be χ < 0.7 for mass ratios Q < 6, with weaker constraints for more compact NSs.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Estimating soybean yields from high-temporal-resolution multi-source data using deep learning

Accurate and timely crop yield prediction is crucial for ensuring food security and maintaining stable agricultural markets. In recent years, there has been a surge in interest in leveraging high-temporal-resolution, multi-source data for effective crop growth monitoring and yield estimation. A notable challenge arises from the difficulty in capturing the intricate interactions between variables across different time steps within these high-temporal-resolution time series datasets. This complexity hinders the reliable extraction of yield information from voluminous and often noisy datasets, especially during periods of extreme weather events. Here, in this study, we propose an Attention and Graph Isomorphism Network-enhanced Bi-directional Long Short-Term Memory network (AGB-LSTM) for estimating county-level soybean yield in the United States. This model integrates a diverse set of remote sensing data, including Near-Infrared Reflectance of Vegetation (NIRv), Sun-Induced chlorophyll Fluorescence (SIF), and Gross Primary Productivity (GPP), along with environmental covariates. The AGB-LSTM effectively leverages information related to crop yield from high-temporal-resolution time series data (5-days), achieving an accuracy of R²= 0.67 and rRMSE = 14.46%. This approach significantly outperforms traditional machine learning methods such as Random Forest (RF) (R²= 0.52, rRMSE = 17.36%) and Bi-LSTM (R²= 0.58, rRMSE = 16.17%). Sensitivity experiments with different time steps and ranges demonstrated that our model could accurately and stably predict yields 1 to 2 months before harvest. Moreover, data with a finer temporal resolution consistently improved prediction performance, resulting in an approximately 20% increase in and an approximately 20% decrease in rRMSE compared to using monthly composites. We also evaluated the robustness of the model under extreme climate events and observed strong performance (R²= 0.50, rRMSE = 21.32%). Finally, yield mapping for major soybean-producing regions in North America in 2023 revealed spatial patterns that closely matched USDA yield reports. Our findings suggest that the AGB-LSTM model is a promising and effective method for estimating yield and has notable potential for global crop yield forecasting.

Deep learning↗
DOE OSTI2023

Learning diffractive optical communication around arbitrary opaque occlusions

Abstract Free-space optical communication becomes challenging when an occlusion blocks the light path. Here, we demonstrate a direct communication scheme, passing optical information around a fully opaque, arbitrarily shaped occlusion that partially or entirely occludes the transmitter’s field-of-view. In this scheme, an electronic neural network encoder and a passive, all-optical diffractive network-based decoder are jointly trained using deep learning to transfer the optical information of interest around the opaque occlusion of an arbitrary shape. Following its training, the encoder-decoder pair can communicate any arbitrary optical information around opaque occlusions, where the information decoding occurs at the speed of light propagation through passive light-matter interactions, with resilience against various unknown changes in the occlusion shape and size. We also validate this framework experimentally in the terahertz spectrum using a 3D-printed diffractive decoder. Scalable for operation in any wavelength regime, this scheme could be particularly useful in emerging high data-rate free-space communication systems.

36 MATERIALS SCIENCE↗