Results for “distributed learning” — Page 14

DOE OSTI2024

Managing changes in peak demand from building and transportation electrification with energy efficiency [Slides]

The Department of Energy funded Berkeley Lab to provide technical assistance to two municipal utilities on how energy efficiency and demand flexibility can mitigate the peak demand impacts of building and transportation electrification. Berkeley Lab worked with these utilities, Sacramento Municipal Utility District (SMUD) and Fort Collins Utilities, to identify research questions that supported their planning needs. For both utilities, Berkeley Lab developed scenario-based load forecasts that considered baseline and high-efficiency building electrification. For SMUD, the forecast also explored the sensitivity of peak demand to extreme weather (a winter cold snap) at the system-level. For Fort Collins Utilities, the forecast addressed the impacts of low, medium, and high levels of building and transportation technology adoption on select distribution feeders. Berkeley Lab is also developing a guidance document for utilities that will draw on lessons learned from the technical assistance and provide a framework for conducting similar analyses.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2026

Real-Time Inference For MI/RR Deblending

The Fermilab Main Injector (MI) and Recycler Ring (RR) share a common beam loss monitor (BLM) system, making loss events difficult to attribute to their source machine when beam is present in both simultaneously. The Real-time Edge AI for Distributed Systems (READS) project addresses this by deblending BLM readings in real time using machine learning (ML). The current FPGA based implementation meets the sub-3 ms latency requirement but carries a resource intensive hls4ml development cycle, motivating exploration of GPU based deployment. This paper characterizes inference latency on an NVIDIA Jetson Orin Nano and introduces a packet organization scheme for assembling synchronized event frames from seven distributed BLM DAQ streams. Using a Python based DAQ simulation with injected timing jitter in place of unavailable live beam data, the pipeline achieved an average end to end latency of 0.456 ms (σ = 0.122 ms) across 167,000 test frames, comfortably meeting the timing constraint. Early outliers were attributed to TensorRT warm-up rather than steady state limitations, suggesting GPU based inference is a viable alternative to the existing FPGA implementation.

Yu, Kellen [Cornell U.]↗

DOE OSTI2026

Exploration of Real Time Inference for MI-RR Deblending on GPU/TPU Systems

The Fermilab Main Injector (MI) and Recycler Ring (RR) share a common beam loss monitor (BLM) system, making loss events difficult to attribute to their source machine when beam is present in both simultaneously. The Real-time Edge AI for Distributed Systems (READS) project addresses this by deblending BLM readings in real time using machine learning (ML). The current FPGA based implementation meets the sub-3 ms latency requirement but carries a resource intensive hls4ml development cycle, motivating exploration of GPU based deployment. This paper characterizes inference latency on an NVIDIA Jetson Orin Nano and introduces a packet organization scheme for assembling synchronized event frames from seven distributed BLM DAQ streams. Using a Python based DAQ simulation with injected timing jitter in place of unavailable live beam data, the pipeline achieved an average end to end latency of 0.456 ms (σ = 0.122 ms) across 167,000 test frames, comfortably meeting the timing constraint. Early outliers were attributed to TensorRT warm-up rather than steady state limitations, suggesting GPU based inference is a viable alternative to the existing FPGA implementation.

Yu, Kellen [Fermilab; Cornell U.]↗

DOE OSTI2025

$\mathrm{SageNet}$: Fast Neural Network Emulation of the Stiff-amplified Gravitational Waves from Inflation

Accurate modeling of the inflationary gravitational waves (GWs) requires time-consuming, iterative numerical integrations of differential equations to take into account their backreaction on the expansion history. To improve computational efficiency while preserving accuracy, we present the Stiff-amplified Gravitational-wave Emulator Network (SageNet), a deep learning framework designed to replace conventional numerical solvers (code available at https://github.com/YifangLuo/SageNet). SageNet employs a long short-term memory architecture to emulate the present-day energy density spectrum of the inflationary GWs with possible stiff amplification, Ω GW (f). Trained on a data set of 25,689 numerically generated solutions, SageNet allows accurate reconstructions of Ω GW (f) and generalizes well to a wide range of cosmological parameters; 90.9% of the test emulations with randomly distributed parameters exhibit errors of under 4%. In addition, SageNet demonstrates its ability to learn and reproduce the artificial, adaptive sampling patterns in numerical calculations, which implement denser sampling of frequencies around changes in spectral indices in Ω GW (f). The dual capability of learning both physical and artificial features of the numerical GW spectra establishes SageNet as a robust alternative to exact numerical methods. Finally, our benchmark tests show that SageNet reduces the computation time from tens of seconds to milliseconds, achieving a speedup of ∼10 4 times over standard CPU-based numerical solvers with the potential for further acceleration on GPU hardware. These capabilities make SageNet a powerful tool for accelerating Bayesian inference procedures for extended cosmological models. In a broad sense, the SageNet framework offers a fast, accurate, and generalizable solution to modeling cosmological observables whose theoretical predictions demand costly differential equation solvers.

Astronomy data modeling↗

DOE OSTI2025

Applying Gaussian Process Machine Learning and Modern Probabilistic Programming to Satellite Data to Infer CO 2 Emissions

Satellite data provides essential insights into the spatiotemporal distribution of CO 2 concentrations. However, many atmospheric inverse models fail to adequately incorporate the spatial and temporal correlations inherent in satellite observations and often lack rigorous methods for estimating parameters like spatial length scales. We introduce an inference model that processes the spatiotemporal covariance in satellite data and estimates hyperparameters such as covariance length scales. Our approach uses the Gaussian process (GP) machine learning (ML) and modern probabilistic programming languages (PPLs) to perform atmospheric inversions of emissions from satellite data. We develop a GP ML inversion system based on modern PPLs and the GEOS-Chem chemical transport model, simulating atmospheric CO 2 concentrations corresponding to the Orbiting Carbon Observatory-2/3 (OCO-2/3) data for July 2020. In our supervised learning framework, we treat the GEOS-Chem simulated data set as the target, with predictors derived by scaling the target with sector-specific factors hidden from the GP machine. Our results show that the GP model, combined with GPU-enabled PPLs, effectively retrieves true emission scaling factors and infers noise levels concealed within the data. This suggests that our method could be applied over larger areas with more complex covariance structures, enabling comprehensive analysis of the spatiotemporal patterns observed in OCO-2/3 and similar satellite data sets.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2024

Experimental Setup and Learning-Based AI Model for Developing Accurate PV Inverter Models

The integration of power electronics-based interfaces presents challenges due to the absence of detailed models and the high computational complexity. Generic models used in system studies lack accuracy in capturing converter dynamics. This paper proposes a data-driven approach developed from experimental setup data. This approach enhances accuracy in photovoltaic inverter modeling. We used two types of PV inverters in the experiment. The recorded experimental data undergo processing through a machine learning model. Results from the model trained through machine learning is also presented.

artificial intelligence↗

DOE OSTI2025

Latent diffusion can map beam loss to two-dimensional phase-space projections

Beam loss monitors (BLMs) and beam current monitors (BCMs) are ubiquitous at particle accelerators around the world. These simple devices provide noninvasive high-level beam measurements but give no insight into the detailed 6D (𝑥,𝑦,𝑧,𝑝 𝑥 ,𝑝 𝑦 ,𝑝 𝑧 ) beam phase-space distributions or dynamics. We show that generative conditional latent diffusion models can learn intricate patterns to solve the extreme inverse problem of mapping waveforms of tens of BLMs or BCMs along an accelerator to detailed 2D projections of a charged particle beam’s 6D phase-space density. This transformational method can be used at any particle accelerator to transform simple noninvasive devices into detailed beam phase-space diagnostics. We demonstrate this concept via multiparticle simulations of the high-intensity beam in the kilometer-long Los Alamos Neutron Science Center linear proton accelerator.

43 PARTICLE ACCELERATORS↗

DOE OSTI2026

An Autonomous MCP Bridge to Rucio: Enhancing Data Management Accessibility for High Energy Physics

The Rucio Data Management System [1] is an important tool used by High Energy Physics experiments, including those at Fermi National Accelerator Laboratory, to store and manage exabyte-scale scientific datasets. Despite its central role in coordinating data across globally distributed storage sites, Rucio's command line interface (CLI) presents a steep learning curve, and makes it difficult for scientists to navigate through. To solve this issue, a containerized Model Context Protocol (MCP) [2] server was built that connects Large Language Models directly to Rucio, allowing AI agents to handle data tasks by using simple, natural language rather than memorized terminal commands. The core engineering focus of this project was moving the server away from slow terminal commands that require text parsing and replacing them with a native Python Client API toolset and a planned REST API framework. Moving to the Python API handles data operations directly in memory, which helps clear up formatting errors, provides the AI with clean, structured JSON data and speeds up tool execution. To prove that the system actually works, a benchmarking pipeline was also built with various questions to test the AI across four different model configurations. The questions included finding data scopes, tracking down specific datasets, and checking replication rules. Through benchmarking, early runs showed that with raw terminal text, the model would get confused and stuck, whereas switching to the Python API to feed the AI clean, structured data yielded massive improvement. By creating an intelligent and autonomous bridge to a storage network, this project shows how AI can be implemented in scientific data management, which ultimately helps scientists at Fermilab spend less time sorting through data and more time focusing on their experiments and analysis.

Akella, Kashyap [William Rainey Harper Coll.]↗

DOE OSTI2026

Adaptive X-ray imaging with reinforcement learning

X-ray imaging is a powerful technique to scan samples in a variety of contexts including biological, environmental and materials science, but commonly requires a synchrotron light source to produce X-rays at sufficient intensity. As these facilities are expensive to operate, the available beam time is limited and always in high demand. Particularly if the illuminated samples are sparse, standard raster scanning methods can be time-consuming, with a majority of that time being spent on areas of the image that carry little information. To increase the efficiency and maximize the information gain for a given time budget, we split the scanning process into a series of steps where previous measurements are used to inform the decision making and adapt the exposure distribution at later stages of the sequence. We formulate this task as a reinforcement learning problem where the goal is to produce a sequence of exposure maps that maximize a predefined scalar metric. We demonstrate the potential of this approach in simulations where the adaptive illumination can accelerate the measurement process by up to an order of magnitude compared with standard raster scanning. Finally, we present the first results from deploying the trained agents on an X-ray fluorescence beamline at the Stanford Synchrotron Radiation Lightsource.

Reinforcement Learning↗

DOE OSTI2024

Establishing and Maintaining Engagement for a Distributional Equity Analysis [Slides]

The Distributional Equity Analysis (DEA) is an analytical framework that allows for the evaluation of the distributional equity of utility resource investments in combination with benefit-cost analysis. This guide serves as a resource for organizations to develop an engagement strategy for a DEA that is accessible, inclusive, and grounded in community interests. It is organized by the guiding principles of engagement for a DEA, summarized in the figure on the right, and offers a compendium of resources for further support and learning.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2025

Different methods of estimating riverbed sediment grain size diverge at the basin scale

Introduction: The distribution of sediment grain size in streams and rivers is often quantified by the median grain size (D50), a key metric for understanding and predicting hydrologic and biogeochemical function of streams and rivers. Manual D50 measurements are time-consuming and ignore larger grains, while approaches to model D50 based on catchment characteristics may over-generalize and miss site-scale heterogeneity. Machine learning-enabled object detection methods like You Only Look Once (YOLO) provides an alternative that enables estimation of D50 that is faster than manual measurements and more site-specific than predictions based on catchment characteristics. Methods: To understand the potential role of object detection methods for improving understanding of D50, we compared D50 estimates made manually, predicted from catchment characteristics, and using a YOLO-enabled approach across the Yakima River Basin. Results: We found distinct differences between methods for D50 averages and variability, and relationships between D50 estimates and basin characteristics. Discussion: We discuss the advantages and limitations of object detection methods versus current methods, and explore potential future directions to combine D50 methods to better estimate spatiotemporal variation of D50, and improve incorporation into basin-scale models.

grain size distribution↗

DOE OSTI2025

Replace Human Intelligence with Fast and Smart Geometric Reasoning and Graph Neural Network to Accelerate Next Gen ModSim Workflows

We present an agent-guided approach to CAD geometry decomposition that automates hex/hybrid meshing with graph neural networks (GNNs) to accelerate next-generation ModSim workflows. Our end-to-end pipeline (i) reduces 3D boundary-representation (B-Rep) models to a 2D chordal axis skeleton (CAT) and then to a 1D bipartite graph of surface and curve nodes, (ii) assigns per node labels as Cubit® WebCut actions, (iii) trains a multi-action GNN under supervised learning, and (iv) predicts five surface-node and three curve-node actions on out-of-distribution test geometries. Each graph node carries geometric, topological, and meshing attributes drawn from the B-Rep “skin” and CAT “skeleton,” with two-way mappings across 3D↔2D↔1D representations to maintain traceability back to 3D CAD. The supervised learning model exhibits stable convergence of the binary cross-entropy loss and achieves 98.7% accuracy on unseen lattice models. To operationalize decision-making, we rank predicted commands by geometric significance and prototyped the agent-guided workflow through the Cubit® Meshing PowerTool GUI. As a stretch goal, we explore reinforcement learning (RL) to reduce or remove label requirements and to learn policies for action sequences that maximize total reward (e.g., size of hex-meshable regions and resulting hex mesh quality). When all-hex meshing is not feasible, the agent assists in producing hybrid meshes—prioritizing hex in critical regions and transitioning to tetrahedral elements (tets) elsewhere—maintaining fidelity while ensuring robustness. The overarching objective is to replace manual, heuristics-based decomposition with data-driven, reproducible automation, cutting meshing turnaround time by orders of magnitude. We anticipate direct impact on simulation workflows through intelligent, scalable decomposition of complex CAD models into hex-meshable subdomains.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Closing the Loop between In Situ Stress Complexity and EGS Fracture Complexity

We present an agent-guided approach to CAD geometry decomposition that automates hex/hybrid meshing with graph neural networks (GNNs) to accelerate next-generation ModSim workflows. Our end-to-end pipeline (i) reduces 3D boundary-representation (B-Rep) models to a 2D chordal axis skeleton (CAT) and then to a 1D bipartite graph of surface and curve nodes, (ii) assigns per node labels as Cubit® WebCut actions, (iii) trains a multi-action GNN under supervised learning, and (iv) predicts five surface-node and three curve-node actions on out-of-distribution test geometries. Each graph node carries geometric, topological, and meshing attributes drawn from the B-Rep “skin” and CAT “skeleton,” with two-way mappings across 3D↔2D↔1D representations to maintain traceability back to 3D CAD. The supervised learning model exhibits stable convergence of the binary cross-entropy loss and achieves 98.7% accuracy on unseen lattice models. To operationalize decision-making, we rank predicted commands by geometric significance and prototyped the agent-guided workflow through the Cubit® Meshing PowerTool GUI. As a stretch goal, we explore reinforcement learning (RL) to reduce or remove label requirements and to learn policies for action sequences that maximize total reward (e.g., size of hex-meshable regions and resulting hex mesh quality). When all-hex meshing is not feasible, the agent assists in producing hybrid meshes—prioritizing hex in critical regions and transitioning to tetrahedral elements (tets) elsewhere—maintaining fidelity while ensuring robustness. The overarching objective is to replace manual, heuristics-based decomposition with data-driven, reproducible automation, cutting meshing turnaround time by orders of magnitude. We anticipate direct impact on simulation workflows through intelligent, scalable decomposition of complex CAD models into hex-meshable subdomains.

42 ENGINEERING↗

DOE OSTI2024

LandScan Global 2023: Silver Edition

For a quarter of a century, the LandScan Global (LSG) project has annually released a global, high-resolution gridded population dataset representing the ambient or unwarned population at a 30 arcsecond resolution. LSG supports a range of applications such as emergency management, disaster response, and human health and security for understanding populations at risk. The 2023 release of LSG, the LandScan Silver Edition, represents a major methodological leap forward while also leveraging previous knowledge—the previous year was the baseline for the current annual update carrying forward valuable knowledge of the built environment for the past quarter century—to train the machine learning models. Compared with annual releases over the past 24years, multiple advancements were made to different aspects of the methodology to achieve reproducibility, transparency, and consistent global propagation of solutions to modeling or population distribution issues identified during the review process. These novel changes include incorporation of the latest available geospatial inputs across the globe, machine learning models instead of manual modifications, population feature importance analysis, open-source solutions vs. proprietary software, generation of multiple global versions, analytic validations, and human-in-the-loop revisions to produce the final version. Additionally, algorithms—such as anomaly detection—were introduced to quickly identify areas of focus to develop a new and robust systematic review. Significant changes in modeled population distributions were observed between the 2022 and 2023 releases, largely attributable to improvements in data and methods and discussed thoroughly within this report. In summation, the LandScan Silver Edition leverages the best of the past quarter century of LSG legacy knowledge and continues a tradition of applying cutting-edge enhancements to serve as a new benchmark for accurate, actionable gridded population data

Lebakula, Viswadeep↗

DOE OSTI2026

Machine learning-based bias-corrected future projections of ozone concentrations from a chemistry-climate model

Reliable projection of future near-surface ozone is crucial for air quality management and health risk assessment. However, potential biases in spatial distribution, magnitude and trends in ozone concentrations simulated by global chemistry-climate models limit their applicability in regional-scale evaluations. In this study, LightGBM, a machine learning (ML) algorithm is applied to correct biases in CESM2-simulated ozone concentrations over China, the United States and Europe and calibrate future ozone projections under two diverse Shared Socioeconomic Pathways (SSP1-2.6 and SSP5-8.5) scenarios from 2020 to 2060. The ML-based correction significantly improves the spatial distribution and reduces the model bias by 40%–60%. It also reverses the potentially incorrect trend of ozone change under SSP1-2.6 in eastern China. When applying ML-based bias correction to CESM2 future projections, warm season mean ozone concentrations decrease across China, the United States, and Europe by –13.5, –17.9, and –13.7 µg/m³, respectively, between 2020 and 2060 in SSP1-2.6, while they increase by 9.4, 2.0, and 5.2 µg/m³ in SSP5-8.5. Decomposition analysis show that changes in anthropogenic emissions dominate future ozone changes in both scenarios, while strong climate penalty from ozone changes occurs in polluted eastern China and climate benefit is found in western China, the United States and Europe under SSP5-8.5. These findings demonstrate the value of combining ML with chemistry-climate models to produce more accurate air quality projections, thereby informing more effective and region-specific environmental protection strategies.

Chemistry Model↗

DOE OSTI2026

Decoding the proton’s gluonic density with lattice QCD-informed machine learning

We present a first machine learning-based decoding of the gluonic structure of the proton from lattice QCD using a variational autoencoder inverse mapper (VAIM). Harnessing the power of generative AI, we predict the parton distribution function (PDF) of the gluon given information on the reduced pseudo-Ioffe-time distributions (RpITDs) as calculated from an ensemble with lattice spacing a ≈ 0.09 fm and a pion mass of M π ≈ 310 MeV. The resulting gluon PDF is consistent with phenomenological global fits within uncertainties, particularly in the intermediate-to-high-x region where lattice data are most constraining. A subsequent correlation analysis confirms that the VAIM learns a meaningful latent representation, highlighting the potential of generative AI to bridge lattice QCD and phenomenological extractions within a unified analysis framework.

Gluon parton distribution function↗

DOE OSTI2025

LLM-Based Adaptive Distribution Voltage Regulation Under Frequent Topology Changes: An In-Context MPC Framework

This paper proposes a large language model (LLM) based adaptive inverter control for distribution voltage regulation under frequent topology changes. We leverage the ability of the LLM to perform in-context learning and create a topology-adaptive surrogate model for power flow calculation. The surrogate model is then integrated with a long short-term memory-based load forecaster and a model predictive control (MPC) scheme to achieve the optimal inverter control that adapts to frequent topology changes. Unlike many existing works that assume fixed-topology grids or require the knowledge of all possible topologies when training a model, the proposed in-context MPC method tackles the distribution voltage control problem under various topologies and adapts to unknown topologies with limited data requirement for fine-tuning. The effectiveness of our method is demonstrated on a modified IEEE 123-bus test system.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2025

Out-of-distribution detection with non-parametric density estimation for models predicting processing history of uranium ore concentrates

The rapid advancement in machine learning (ML) and computer vision (CV) coincides with the growth of interest in deploying these ML/CV models in numerous fields from medicine to social science. Similar to those areas, we have witnessed a great number of works in materials science employing ML/CV models – neural networks in particular – in their studies in recent years. These models have proven to obtain accurate performance in various tasks. However, these models struggle to attain a similar performance when encountering test samples coming from a distribution that is different from the training set. More importantly, they fail without providing any warning to the users. Therefore, we propose a framework for detecting out-of-distribution (OOD) samples to alert users when a human intervention might be necessary in this work. Specifically, we explore the use of a non-parametric density estimation method to detect OOD samples. Here, we assess OOD detection capability of the proposed framework on ML models developed for categorizing precipitation routes of U 3 O 8 when encountering OOD datasets that contain samples (1) undergone different imaging acquisition process, (2) undergone different material synthesis process, and (3) different materials than ID set. Through those experiments, we achieve an average area under the receiver operating characteristic (AUROC) of at least 91% on average in detecting OOD samples. With minimal overhead cost and superior performance, the proposed framework enables a reliable and safe system when deploying in real-world scenarios.

Convolutional neural networks↗