Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “network acceleration”

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 1,135 records · Page 63
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Constraining the destruction rate of K 40 in stellar nucleosynthesis through the study of the Ar 40 ( p , n ) K 40 reaction

Background: We present that K 40 plays a significant role in the radiogenic heating of Earth-like exoplanets, which can affect the development of a habitable environment on their surfaces. The initial amount of K 40 in the interior of these planets depends on the composition of the interstellar clouds from which they formed. Within this context, nuclear reactions that regulate the production of K 40 during stellar evolution can play a critical role. Purpose: In this study, we constrain for the first time the astrophysical reaction rate of K 40 ( n , p ) Ar 40 , which is responsible for the destruction of K 40 during stellar nucleosynthesis. We provide to the nuclear physics community high-resolution data on the cross section and angular distribution of the Ar 40 ( p , n ) K 40 reaction. These are important to various applications involving Ar 40 . The associated reaction rate of the Ar 40 ( p , n ) K 40 process addresses a reaction rate gap in the Joint Institute for Nuclear Astrophysics REACLIB database in the region of intermediate-mass isotopes. Methods: We performed differential cross-section measurements on the Ar 40 ( p , n ) K 40 reaction, for six energies in the center-of-mass system between 3.2 and 4.0 MeV and various angles between 0 ° and 135 ° . The experiment took place at the Edwards Accelerator Laboratory at Ohio University using the beam swinger target location and a standard neutron time-of-flight technique. We extracted total and partial cross sections by integrating the double differential cross sections we measured. Results: The total and partial cross sections varied with energy due to the contribution from isobaric analog states and Ericson type fluctuations. The energy-averaged neutron angular distributions were symmetrical relative to 90 ° . Based on the experimental data, local transmission coefficients were extracted and were used to calculate the astrophysical reaction rates of Ar 40 ( p , n ) K 40 and K 40 ( n , p ) Ar 40 reactions. The new rates were found to vary significantly from the theoretical rates in the REACLIB library. We implemented the new rates in network calculations to study nucleosynthesis via the slow neutron capture process, and we found that the produced abundance of K 40 is reduced by up to 10% compared to calculations with the library rates. At the same time, the above result removes a significant portion of the previous theoretical uncertainty on the K 40 yields from stellar evolution calculations. Conclusions: Our results support a destruction rate of K 40 in massive stars via the K 40 ( n , p ) Ar 40 reaction that is larger compared to previous estimates. The rate of K 40 destruction via the K 40 ( n , p ) Ar 40 reaction now has a dramatically reduced uncertainty based on our measurement. Lastly, this result directly affects the predicted stellar yields of K 40 from nucleosynthesis, which is a critical input parameter for the galactic chemical evolution models that are currently employed for the study of significant properties of exoplanets.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS1996

Hypersonic Airbreathing Vehicles/Technologies

Hypersonic airbreathing horizontal takeoff and landing (HTOL) vehicles are highly integrated systems involving many advanced technologies. The design environment is variable rich, intricately networked, and sensitivity intensive; as such, it represents a tremendous challenge. Creating a viable design requires addressing three main elements: (1) an understanding of the 'figures of merit' and their relationship, (2) the development of sophisticated configuration discipline prediction methods and a synthesis procedure, and (3) the synergistic integration of advanced technologies across the discipline spectrum. This paper will focus on the vision for hypersonic airbreathing vehicles and the advanced technologies that forge the designs. Airbreathing hypersonics encompass endoatmospheric (airplanes...missiles are a part of the matrix but will not be included in this paper since they are an air force focus) and space access vehicles with speed from Mach 4 up to Mach 25 (orbital). These vehicles can be divided into two classes...cruisers and accelerators. The cruiser designs reflect high lift-to-drag whereas the accelerators reflect low drag per unit inlet capture; thus, the cross section of the accelerator attributes a much larger percentage to propulsion. One of the more design influencing items is fuel. The hydrogen fueled vehicles must be very volumetric efficient to contain the low density fuel and thus tend to be a bit bulgy (more conducive to lifting bodies or wing bodies) whereas with hydrocarbon fueled vehicles, the concern is loading because of the high density fuel; thus, they may tend to be more towards waveriders which are not usually very volumetric efficient. Hydrocarbon fuels (endothermic) are limited in their engine cooling capacity to below Mach 8.

Hunt, James L.↗
DOE OSTI2025

In-Field Testing of Components for Feedback and Control of the ITER Disruption Mitigation System

Here, the shattered pellet injection (SPI) method has been chosen as the disruption mitigation system (DMS) for ITER. To protect the device from plasma disruptions that cause damaging heat and electromagnetic loads, SPI is used to inject high-Z material into the plasma. The process of SPI utilizes cryogenic cooling to form solid pellets. Pellets are accelerated down a barrel and into an angled surface, causing the pellet to shatter prior to entering the tokamak chamber. For the DMS to function reliably, the 27 separate shattered pellet injectors planned for ITER must rely on many components to provide accurate feedback data and for control functions. Each component in the DMS is exposed to an elevated background magnetic field depending on its placement and proximity to the plasma chamber. A Helmholtz coil test stand that is operated at Oak Ridge National Laboratory was utilized to test the components in relevant background field levels to assess component performance. This paper details the test design and results for in-field component operation for a variety of components. This list includes the following components: two different network switches for camera connectivity, a VAT fast shutter valve intended to reduce the flow of SPI propellant gas into the torus, a solenoid control valve intended for use in the pellet formation process, pressure/vacuum switches to be used for feedback and control, a printed circuit board piezo pressure sensor to be used to measure breech pressure, and various relays for the high-voltage pulsed power supply used to drive the SPI propellant valve.

Disruption mitigation↗
DOE OSTI2023

In planta single-molecule imaging and holographic force spectroscopy to study real-time, multimodal turnover dynamics of polysaccharides and associated carbohydrate metabolites

Cellular production of polysaccharides from simple sugar molecules serves critical roles in a variety of metabolic processes essential to the survival of every living organism. In the context of photosynthetically fixed carbon, cell wall polysaccharides synthesis is particularly of interest for increased biomass accumulation for human food, animal feed, and bioenergy related applications. Despite a long history of research in plant cell biology, our understanding of in planta cell wall biosynthesis and its regulation is far from complete due to lack of a comprehensive microscopy toolkit that encompass the multiple spatial and temporal scales in which cell wall polysaccharides fibrils are synthesized and assembled into intricate cell wall networks. In this project, a multidisciplinary team of scientists from Rutgers University, Vanderbilt University, and Oak Ridge National Laboratory carried out innovative multimodal, multiscale microscopy studies of cell wall synthesis using single-molecule force-spectroscopy, super-resolution fluorescence microscopy, and in vivo live cell imaging. Our novel multimodal and holistic imaging approach revealed plant cell wall polysaccharides synthesis processes in unprecedented detail across multiple spatiotemporal scales, from a single-molecule to a single-cell. Besides, the team has made technical and scientific innovations across multiple research fields—microscopy, bioengineering, single-molecule biophysics, and plant biology, etc.—to accomplish the goals. The results from this project will greatly advance the mechanistic and holistic understanding of in planta cell wall synthesis, which will accelerate the development of better transgenic crops for bioenergy applications. Moreover, the new toolbox, combining powerful advanced microscopy assays with cell/protein engineering, will have broader impacts on molecular and cellular biology fields by paving the way for studying cellular processes occurring across multiple physical scales with multimodal microscopy methodologies.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Delivery Ring Spill Characterization and Impulse Study

High-intensity particle physics experiments require uniform beam extraction to prevent instantaneous rate spikes from overwhelming detector systems. By analyzing accelerator parameters and extracted beam dynamics, we directly inform spill regulation systems that make real-time adjustments to minimize non-uniformity. This Department of Energy Visiting Faculty Program project transitioned from characterizing Main Injector half-integer slow extraction for SpinQuest to Delivery Ring third-integer slow extraction for Mu2e. Working alongside the Fast Adaptive Neural Control (FANC) group, we developed an automated pipeline that aligns asynchronous instrument channels, embeds quality metrics, and isolates clean spill populations. Analyzing baseline spills alongside a dedicated quadrupole impulse study allowed us to quantify noise structures while mapping time-varying beam response and transit-delay dynamics. These empirical measurements directly ground digital twin models, supporting FANC’s deployment of real-time, FPGA-based neural network controllers in the Mu2e Spill Regulation System.

Dolen, James William [Purdue U., West Lafayette] (↗
DOE OSTI2025

Surrogate Neural Architecture Codesign Package (SNAC-Pack)

Neural Architecture Search is a powerful approach for automating model design, but existing methods struggle to accurately optimize for real hardware performance, often relying on proxy metrics such as bit operations. We present Surrogate Neural Architecture Codesign Package (SNAC-Pack), an integrated framework that automates the discovery and optimization of neural networks focusing on FPGA deployment. SNAC-Pack combines Neural Architecture Codesign's multi-stage search capabilities with the Resource Utilization and Latency Estimator, enabling multi-objective optimization across accuracy, FPGA resource utilization, and latency without requiring time-intensive synthesis for each candidate model. We demonstrate SNAC-Pack on a high energy physics jet classification task, achieving 63.84% accuracy with resource estimation. When synthesized on a Xilinx Virtex UltraScale+ VU13P FPGA, the SNAC-Pack model matches baseline accuracy while maintaining comparable resource utilization to models optimized using traditional BOPs metrics. This work demonstrates the potential of hardware-aware neural architecture search for resource-constrained deployments and provides an open-source framework for automating the design of efficient FPGA-accelerated models.

Weitz, Jason [UC, San Diego] (ORCID:00090004631535↗
DOE OSTI2024

Cultural Shifts in High Energy Physics Collaboration from the Cold War to the Present: A Historical and Philosophical Perspective

Here, this article employs empirical history and the philosophy of science to study cultural convergences and divergences in international collaborations in high energy physics. We examine two cases: (1) E-36, an experiment on small angle proton-proton scattering conducted during the Cold War at the National Accelerator Laboratory (NAL) in the USA by Soviet and US scientists and (2) an ongoing collaborative experiment, NICA, at the Joint Institute for Nuclear Research (JINR, Dubna), which is a project devoted to heavy-ion physics. The JINR, particularly its Laboratory of High Energy Physics (formerly the “Laboratory of High Energies”) is the main mediating actor between these two cases (i.e., E-36 and NICA), as the majority of Soviet participants in E-36 were representatives of the Institute. Using empirical data collected through archival searches, field observations conducted at JINR in 2018–2019, and in-depth interviews, we tell a story of cultural differences in high energy physics by applying the concepts of ‘trading zones’ (P. Galison) and the translation of interests in actor-networks (B. Latour, M. Callon and others). We analyze three types of cultural diversity (specialization, nationality, and generational) in light of the implications of temporal context and the dichotomy between East and West, showing the roles cultural diversity plays in scientific collaboration (which is an integral part of as well as obstacle to scientific research that can nevertheless provide learning opportunities). Our study aims to demonstrate how disunity and diversity may function in scientific research and how high energy physics collaborations can remain productive despite sometimes deep divergences, including those between East and West.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2023

Accelerating Scientific Simulations with Bi-Fidelity Weighted Transfer Learning

High-fidelity modeling is an essential design tool for many engineering applications. However, for complex systems, computational cost can be a limiting factor. Analyzing parameter sensitivity, uncertainty quantification, and design optimization require many model evaluations. Surrogate models are often used to develop the relationship between model parameters and quantities of interest. However, in the case of complex systems, surrogate models require several degrees of freedom and, thus, a large number of data points to determine the correct dependencies. For many applications, this may be prohibitively expensive. The reduction of computational requirements can be achieved by leveraging low-fidelity models. Low-fidelity models represent the system at a coarser resolution with the advantage of computational efficiency. Therefore, a bi-fidelity modeling paradigm, which augments the accuracy of a low-fidelity model in a computationally efficient manner by invoking limited runs of a high-fidelity model, can be leveraged to sufficiently balance the accuracy and computational requirements. In this work, a bi-fidelity weighted transfer learning method using neural networks was applied to a computational fluid dynamics heat transfer modeling problem. The transfer learning advantage was investigated as a function of hyperparameters. Our main finding is that the use of a bi-fidelity modeling paradigm achieves accuracy close to that of a high-fidelity Gaussian process model while significantly reducing computational cost. The bi-fidelity model achieves comparable performance with 90 high-fidelity samples-that is, 60% less than the samples needed to achieve similar accuracy without the use of bi-fidelity modeling,

Borowiec, Katarzyna↗
DOE OSTI2024

Smart pixel sensors Towards on-sensor filtering of pixel clusters with deep learning

High granularity silicon pixel sensors are at the heart of energy frontier particle physics collider experiments. At an collision rate of 40\,MHz, these detectors create massive amounts of data. Signal processing that handles data incoming at those rate and intelligently reduces the data within the pixelated region of the detector \textit{at rate} will enhance physics performance and enable physics analyses that are not currently possible. Using the shape of charge clusters deposited in an array of small pixels, the physical properties of the traversing particle can be extracted with locally customized neural networks. In this first work, we present a neural network that can be embedded into the on-sensor readout and filter out hits from low momentum tracks, reducing the detector's data volume by 54.4-75.4\%. The network is designed and simulated as a custom readout integrated circuit with 28\,nm CMOS technology and is expected to operate at less than 300\,$\mu W$ with an area of less than 0.2\,mm$^2$.

43 PARTICLE ACCELERATORS↗
DOE OSTI2025

IoT-based retrofit information diffusion in future smart communities

Community-scale building retrofits are not merely scaled-up versions of single-building retrofits. They involve complex challenges, such as reconciling individual interests with collective goals and managing the dynamic interplay between buildings through mechanisms like power grids and social connections. Internet of Things (IoT) connectivity holds the potential to leverage these interplays to balance individual and collective interests effectively in smart communities. One critical aspect of this interplay is information diffusion, which shapes how retrofit decisions spread among neighbors, influencing individual choices and ultimately impacting community-level retrofit outcomes. In other words, IoT-based smart devices automatically push tailored retrofit notifications to homeowners, which completely changes the format of information diffusion in the future. To investigate this influence by such information diffusion, the study used CityBES to simulate energy performance for different retrofits and applied an information diffusion model to analyze how decisions spread in a networked community of 192 buildings. The diffusion process was modeled on a weighted, directed network, capturing the dynamics of information flow and decision-making across 16 scenarios. Individual retrofit benefits were evaluated through payback years, while community-level retrofit outcomes were assessed using greenhouse gas (GHG) emission reductions. The results demonstrate that easier information diffusion among neighbors encourages households to prioritize retrofit measures that align with the majority’s optimal choices, even at the expense of individual financial benefits. In this case, such collective prioritization enhanced community-level retrofit performance, increasing GHG emission reductions by up to 29.4 %. However, this improvement came with trade-offs, as the average payback period for households extended by approximately 1.74 years. These findings highlight the potential of IoT-based information diffusion in future smart communities to coordinate individual interests with collective goals, ultimately accelerating community-level building retrofits.

Shu, Lei↗
NASA NTRS1999

Gravitational Acceleration Effects on Macrosegregation: Experiment and Computational Modeling

Experiments were performed under terrestrial gravity (1g) and during parabolic flights (10-2 g) to study the solidification and macrosegregation patterns of Al-Cu alloys. Alloys having 2% and 5% Cu were solidified against a chill at two different cooling rates. Microscopic and Electron Microprobe characterization was used to produce microstructural and macrosegregation maps. In all cases positive segregation occurred next to the chill because shrinkage flow, as expected. This positive segregation was higher in the low-g samples, apparently because of the higher heat transfer coefficient. A 2-D computational model was used to explain the experimental results. The continuum formulation was employed to describe the macroscopic transports of mass, energy, and momentum, associated with the solidification phenomena, for a two-phase system. The model considers that liquid flow is driven by thermal and solutal buoyancy, and by solidification shrinkage. The solidification event was divided into two stages. In the first one, the liquid containing freely moving equiaxed grains was described through the relative viscosity concept. In the second stage, when a fixed dendritic network was formed after dendritic coherency, the mushy zone was treated as a porous medium. The macrosegregation maps and the cooling curves obtained during experiments were used for validation of the solidification and segregation model. The model can explain the solidification and macrosegregation patterns and the differences between low- and high-gravity results.

Leon-Torres, J.↗
DOE OSTI2026

Image processing pipeline for AI-driven nanoparticle megalibrary characterization

Recent innovations have made it possible to produce megalibraries, millions of structurally and compositionally distinct nanoparticles on a chip. These megalibraries yield vast volumes of data that are impossible to analyze manually, necessitating the development of automated tools. In previous work, we created a binary classification machine learning model to select quality nanoparticle images for downstream analysis. In this work, we show that adding a custom image processing step before training can produce significantly higher-performing models in a fraction of the time and make them more robust to different image noise levels and microscope acquisition settings. The image processing pipeline proposed here effectively cleans raw nanoparticle images, enhances key features, and allows us to use much lower resolution images and simpler neural network model architectures. These features result in higher performance and significant cost savings. Experiments demonstrate superior performance relative to baseline, including an 18.2% improvement in recall and a 13.1% increase in accuracy. Given the high cost of downstream analysis, it is critical to minimize false positives, and our best-performing model reaches a precision of 95.9% and a weighted F-score of 95.1% on an unseen test set. Additionally, model training time is reduced from hours to less than a minute. We also show that, using this custom image processing pipeline, model performance is significantly improved at lower pixel resolutions compared to downsizing alone. We expect that adopting this pipeline for AI-driven automated nanoparticle characterization will allow researchers to rapidly and accurately analyze much greater volumes of data, thereby accelerating materials discovery.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2024

Two-Step Hyperparameter Optimization Method: Accelerating Hyperparameter Search by Using a Fraction of a Training Dataset

Abstract Hyperparameter optimization (HPO) is an important step in machine learning (ML) model development, but common practices are archaic—primarily relying on manual or grid searches. This is partly because adopting advanced HPO algorithms introduces added complexity to the workflow, leading to longer computation times. This poses a notable challenge to ML applications, as suboptimal hyperparameter selections curtail the potential of ML model performance, ultimately obstructing the full exploitation of ML techniques. In this article, we present a two-step HPO method as a strategic solution to curbing computational demands and wait times, gleaned from practical experiences in applied ML parameterization work. The initial phase involves a preliminary evaluation of hyperparameters on a small subset of the training dataset, followed by a reevaluation of the top-performing candidate models postretraining with the entire training dataset. This two-step HPO method is universally applicable across HPO search algorithms, and we argue it has attractive efficiency gains. As a case study, we present our recent application of the two-step HPO method to the development of neural network emulators for aerosol activation. Although our primary use case is a data-rich limit with many millions of samples, we also find that using up to 0.0025% of the data—a few thousand samples—in the initial step is sufficient to find optimal hyperparameter configurations from much more extensive sampling, achieving up to 135× speedup. The benefits of this method materialize through an assessment of hyperparameters and model performance, revealing the minimal model complexity required to achieve the best performance. The assortment of top-performing models harvested from the HPO process allows us to choose a high-performing model with a low inference cost for efficient use in global climate models (GCMs).

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1993

The coming technological singularity: How to survive in the post-human era

The acceleration of technological progress has been the central feature of this century. I argue in this paper that we are on the edge of change comparable to the rise of human life on Earth. The precise cause of this change is the imminent creation by technology of entities with greater than human intelligence. There are several means by which science may achieve this breakthrough (and this is another reason for having confidence that the event will occur): (1) the development of computers that are 'awake' and superhumanly intelligent (to date, most controversy in the area of AI relates to whether we can create human equivalence in a machine. But if the answer is 'yes, we can', then there is little doubt that beings more intelligent can be constructed shortly thereafter); (2) large computer networks (and their associated users) may 'wake up' as a superhumanly intelligent entity; (3) computer/human interfaces may become so intimate that users may reasonably be considered superhumanly intelligent; and (4) biological science may find ways to improve upon the natural human intellect. The first three possibilities depend in large part on improvements in computer hardware. Progress in computer hardware has followed an amazingly steady curve in the last few decades. Based largely on this trend, I believe that the creation of greater than human intelligence will occur during the next thirty years.

Vinge, Vernor↗
DOE OSTI2020

Lighting Technology Energy Solutions (LiTES) (Final Technical Report)

The Lighting Technologies Energy Solutions (LiTES) Program Final Technical Report is a collaborative effort between NextEnergy, Consumers Energy, DTE Energy (DTE) and the Detroit Joint Apprentice Training Center (DJATC) to evaluate customer and contractor experiences deploying advanced/networked lighting controls systems (ALC/NLC). The LiTES Program, funded by the US Department of Energy (DOE), sought to reduce energy use in small and medium commercial buildings (SMCB) by accelerating the adoption of ALC/NLC through contractor training and technology deployment. Leveraging recommendations already outlined by the Design Lights Consortium Commercial Advanced Lighting Controls (DLC CALC) project, NextEnergy led an effort to train contractors, evaluate the experience of ALC/NLC demonstration projects, identify opportunities to reduce market barriers, and accelerate the increased adoption of ALC/NLC technologies within SMCB. The LiTES Program defined SMCB as commercial buildings under 100,000 square feet. The LiTES Program efforts also included evaluating current ALC/NLC utility incentives, piloting ALC/NLC incentives specific to SMCB, and identifying opportunities to better align utility incentives with current ALC/NLC technology to support accelerating the adoption of ALC/NLC in SMCB. This report shares the survey results and M&V data findings, including energy savings, market barriers, experience of installations, impact of incentives/financing programs and provides recommendations for the increased deployment of ALC/NLC systems in small and medium commercial buildings. Utility survey results, incentive alternatives and recommendations, including a cost benefit analysis outlining energy and economic impacts of the provided ALC/NLC incentives on SMCB projects are also provided.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2024

Physics-informed graphical neural network for power system state estimation

State estimation is highly critical for accurately observing the dynamic behavior of the power grids and minimizing risks from cyber threats. However, existing state estimation methods encounter challenges in accurately capturing power system dynamics, primarily because of limitations in encoding the grid topology and sparse measurements. Here, this paper proposes a physics-informed graphical learning state estimation method to address these limitations by leveraging both domain physical knowledge and a graph neural network (GNN). We employ a GNN architecture that can handle the graph-structured data of power systems more effectively than traditional data-driven methods. The physics-based knowledge is constructed from the branch current formulation, making the approach adaptable to both transmission and distribution systems. The validation results of three IEEE test systems show that the proposed method can achieve lower mean square error more than 20% than the conventional methods.

43 PARTICLE ACCELERATORS↗
DOE OSTI2025

Building molecular model series from heterogeneous CryoEM structures using Gaussian mixture models and deep neural networks

Cryogenic electron microscopy (CryoEM) produces structures of macromolecules at near-atomic resolution. However, building molecular models with good stereochemical geometry from those structures can be challenging and time-consuming, especially when many structures are obtained from datasets with conformational heterogeneity. Here we present a model refinement protocol that automatically generates series of molecular models from CryoEM datasets, which describe the dynamics of the macromolecular system and have near-perfect geometry scores. This method makes it easier to interpret the movement of the protein complex from heterogeneity analysis and to compare the structural dynamics observed from CryoEM data with results from other experimental and simulation techniques.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Fed-DeepONet: Stochastic Gradient-Based Federated Training of Deep Operator Networks

The Deep Operator Network (DeepONet) framework is a different class of neural network architecture that one trains to learn nonlinear operators, i.e., mappings between infinite-dimensional spaces. Traditionally, DeepONets are trained using a centralized strategy that requires transferring the training data to a centralized location. Such a strategy, however, limits our ability to secure data privacy or use high-performance distributed/parallel computing platforms. To alleviate such limitations, in this paper, we study the federated training of DeepONets for the first time. That is, we develop a framework, which we refer to as Fed-DeepONet, that allows multiple clients to train DeepONets collaboratively under the coordination of a centralized server. To achieve Fed-DeepONets, we propose an efficient stochastic gradient-based algorithm that enables the distributed optimization of the DeepONet parameters by averaging first-order estimates of the DeepONet loss gradient. Then, to accelerate the training convergence of Fed-DeepONets, we propose a moment-enhanced (i.e., adaptive) stochastic gradient-based strategy. Finally, we verify the performance of Fed-DeepONet by learning, for different configurations of the number of clients and fractions of available clients, (i) the solution operator of a gravity pendulum and (ii) the dynamic response of a parametric library of pendulums.

Moya, Christian↗