Search NASA⌕ Search

SEARCH · Search NASA

Results for “REDUCING AGENTS”

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 343 records · Page 19
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Design and Fabrication of Polyisocyanurate Foams toward Significantly Enhanced Thermal Resistivity

Developing high-performance thermal insulation is vital for addressing the ongoing global demand for reduced energy costs. Polyisocyanurate (PIR) foams are commonly used in residential and commercial buildings and in various industrial applications owing to their relatively high thermal insulation properties and fire resistivity. Here, this study aims to further improve the thermal resistivity of PIR foams by (1) incorporating low thermal conductivity blowing agents; (2) tailoring the anisotropy of their cells; (3) tuning polymeric isocyanate quantities; and (4) incorporating a facer barrier, while using steps that easily integrate into current manufacturing processes for PIR foams. The resulting PIR foams exhibit a significant enhancement in thermal resistivity, achieving initial values as high as 8.3 h·ft 2 ·°F/Btu/in., commonly abbreviated as R-8.3/in., which is a 20% improvement compared with that of commercially used PIR foams that achieve approximately R-7/in. The detailed analysis of thermal conductivity measurements, mechanical testing, and morphological characterization elucidates the structure–property relationships. The developed high-performance PIR foams provide a critical pillar for next-generation high-performance insulation, offering promising thermal insulation for buildings and many other applications that have a significant effect on global energy costs.

Anisotropic pores↗
DOE OSTI2022

Efficient monolithic all-perovskite tandem solar modules with small cell-to-module derate

All-perovskite tandem solar modules are promising to reduce the cost of photovoltaic systems with their high efficiency and solution fabrication, but their sensitivity to air still imposes a great challenge. Here a hot gas-assisted blading method is developed to accelerate the perovskite solidification, forming compact and thick narrow bandgap (NBG) perovskite films. Adding a reduction agent into NBG films followed by a short period of air exposure and a post-fabrication storage surprisingly increases carrier recombination lifetime and enables laser scribing in ambient conditions without obvious loss of device performance. This combination suppresses tin and iodide oxidation and forms a thin SnO 2 layer on the NBG film surface. Monolithic all-perovskite tandem solar modules showed a champion efficiency of 21.6% with a 14.3 cm 2 aperture area, corresponding to an active area efficiency of 23.0%. Lastly, the very small cell-to-module derate of 6.5% demonstrates the advantage of a tandem monolithic structure for solar modules.

14 SOLAR ENERGY↗
DOE OSTI2024

Wholesale Electricity Markets and Resource Adequacy with High Clean Energy Generation Targets

Wholesale electricity markets are intended to incentivize system generation investments and operations outcomes that meet evolving system needs. In this work, we evaluate the effectiveness of wholesale market structures, rules and policies in achieving system resource adequacy (RA) and clean energy targets in the presence of self-interested generation investors using the Electricity Markets and Investment Suite Agent-based Simulation (EMIS-AS) model. Results highlight that both capacity markets and operating reserve demand curves (ORDCs) can help achieve a reliable system but with different RA compliance timelines and distribution of generation technologies. Structures with capacity markets tend to favor more capital-intensive peaking technologies while reducing wind and solar build-outs due to suppressed energy and clean energy market prices, particularly in the absence of strong clean energy targets. Conversely, ORDCs improve the commitment of available generation units, but this comes at the expense of higher system costs and renewable generation curtailment. We also find that well-calibrated static capacity demand curves can yield similar reliability and total cost compared to capacity market demand curves informed dynamically by resource adequacy while also yielding stable annual capacity prices. Different approaches to formulating ORDC curves can also yield key trade-offs, namely that a more efficient treatment of storage chronology results in lower ORDC curves and prices, yielding less investment and cost but at the expense of reliability. Finally, the effectiveness of wholesale electricity markets in practically achieving very high clean energy generation targets highly depends on the cost-competitiveness of clean energy technologies that can support critical balancing needs across multiple timescales.

capacity expansion↗
DOE OSTI2025

Zooming in on virtual commutes: Telecommuting impacts on mobility and sustainability

Motivated by a societal shift towards remote work and rapid advancement in information and communication technologies, this study examines the impact of telecommuting on urban mobility across the nine counties of the San Francisco Bay Area, California. Utilizing a behaviorally realistic integrated agent-based transportation model and recent data on telecommuting patterns, we simulate the impact of multiple telecommuting scenarios on transportation system outcomes. This provides a comprehensive picture of the impact of remote work on travel costs and accessibility of all travelers, factoring in changes in mode use and congestion. We analyze how telecommuting influences broader societal factors such as transportation energy consumption. Our findings indicate that increased telecommuting reduces overall person miles traveled and transportation energy consumption. Furthermore, telecommuting results in externality benefits by improving accessibility and reducing commute times for non-telecommuters.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2026

Agentic framework for programmatic crystal structure generation using a fine-tuned worker–supervisor large language model

Platinum group metals (PGMs) underpin many catalytic technologies but face severe supply constraints, motivating the search for alternative materials and computational methods to accelerate discovery. While atomistic simulation tools such as Pymatgen and ASE have streamlined structure manipulation, they require detailed inputs, limiting accessibility for experimentalists and slowing early-stage exploration. Here, in this study, we present an AI-driven agentic framework that orchestrates worker–supervisor large language models (LLMs). The worker translates natural-language prompts of varying abstraction into valid crystallographic structures using a compact LLM fine-tuned with low-rank adaptation on a curated text–code–CIF dataset, emphasizing energy-efficient training. Benchmarking against the baseline CodeGen-350M-mono model shows that fine-tuning reduces hallucination rates from 100% to as low as 5% and improves structural match accuracy to up to 82% for fully specified inputs. Accuracy declines with decreasing prompt detail but remains nontrivial even when only stoichiometry and space group are provided, underscoring the LLM’s capacity for crystallographic inference. The supervisor Claude LLM evaluates the outputs and triggers iterative refinement through the worker’s built-in structure manipulation capabilities (e.g., supercell scaling, strain, vacancy, and substitution operations). We further demonstrate use cases for technologically relevant catalysts, including IrO 2 , pyrochlore Pb 2 Ir 2 O 7 , Ni 2 FeO 4 , and Ni 3 Mo, where the framework generates physically consistent structures that can be refined via geometry optimization. This work introduces a low-energy, language-driven pathway for integrating human and machine intelligence in materials design, paving the way for AI-assisted synthesis planning and high-throughput screening of complex oxides.

AI agent↗
DOE OSTI2025

Simulating nationwide coupled disease and fear spread in an agent-based model

Human cognitive responses, behavioral responses, and disease dynamics co-evolve over the course of any disease outbreak, and can result in complex feedbacks. We present a dynamic agent-based model that explicitly couples the spread of disease with the spread of fear surrounding the disease, implemented within the EpiCast simulation framework. EpiCast models transmission within a realistic synthetic population, capturing individual-level interactions. In our model, fear propagates through both in-person contact and broadcast media, prompting individuals to adopt protective behaviors that reduce disease spread. In order to better understand these coupled dynamics, we create and compare a range of compartmental models to ensure that introducing additional disease states does not prevent the emergence of multiple waves in these simpler models. Additionally, we compare a range of behavioral scenarios within EpiCast, varying the level and intensity of fear and behavior change. Our results show that the addition of asymptomatic, exposed, and pre-symptomatic disease states can impact both the rate at which an outbreak progresses and its overall trajectory in compartmental models. In EpiCast, the combination of non-local fear spread via broadcasters and strong behavioral responses by fearful individuals generally leads to multiple epidemic waves, an outcome that occurs only within a narrow parameter range when fear spreads purely through local contact. Accounting for the coupled spread of fear and disease is critical for understanding disease dynamics and designing timely, targeted responses to emerging infectious threats.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

AlphaBuilding ResCommunity: A multi-agent virtual testbed for community-level load coordination

Training and validating algorithms in a simulation testbed can accelerate research and applications of optimal control of residential loads to improve energy flexibility and grid resilience. We developed an open-source simulation environment, AlphaBuilding ResCommunity, that can be used to train and validate algorithms to control a single thermostatically controlled load (TCL) or coordinate a group of TCLs. We used reduced-order models to simulate the thermodynamics of TCLs, and the parameter values were determined from the connected smart thermostat data of real households. The environment was built upon the standardized OpenAI Gym interface. Ancillary functions, such as retrieving the parameters and weather forecasts, are provided to facilitate control strategies that require predictive information. Compared with existing efforts, AlphaBuilding ResCommunity has three advantages: (1) more realistic model settings because the parameter values are identified from actual household operating data, and modelling and measurement uncertainty are considered; (2) passive thermal storage control; and (3) ease of use due to a simple software dependency and standardized interface. We demonstrated the applications of the environment by implementing a Kalman Filter and Model Predictive Control on a single TCL and a Priority-Stack-Based Control and Alternating Direction Method of Multipliers to coordinate multiple TCLs for load tracking.

Wang, Z↗
DOE OSTI2021

Hamilton: Flexible, Open Source $10 Wireless Sensor System for Energy Efficient Building Operation

Sensors for improving building performance are rapidly populating the market, driven in part by the drive to reduce greenhouse gas emissions resulting from energy production as well as improve the interior environment for healthy and more productive spaces. UC Berkeley has led wireless sensor development over the past 25 years (e.g., Telos mote), with the Hamilton (named after Alexander Hamilton on the US $10 bill) as the most recent. The Hamilton sensor was designed as a low-cost high-performance sensor that is modular and interoperable. The objective of the Hamilton project was to create, evaluate and establish the technological foundations for secure and easy to deploy building energy efficiency applications utilizing pervasive, low-cost wireless sensors integrated with traditional Building Management Systems (BMS), consumer-sector building components, and powerful data analytics. The project included iterative hardware design, incorporating a high-performance database (BTrDb, http://btrdb.io/), creating and iterating the development of secure data middleware (BOSSwave, WAVE/WAVEMQ), working with and pushing the development of an open-source tiny operating system RiotOS, and implementing and improving protocols such as Thread/OpenThread and TCP/IP. The hardware benefited from careful design to drive down the cost; the design included a System-on-a-Chip (SoC), chip antenna, single crystal and five passive components. Careful design of the operating system created a low-power design to enable a long life with small batteries. The hardware included several sensors: temperature, radiant temperature, relative humidity, magnetometer, accelerometer, and light, with an optional occupancy (Passive InfraRed) sensor. The project was the basis of several applications, both internal to the research team and other researchers and professionals at other institutions. Several applications used the sensor hardware as the basis for other complex devices. Other applications used the sensors to improve building performance through interoperating with the building Heating Ventilation and Air-Conditioning (HVAC) system, such as using occupancy and/or distributed temperature sensing to reduce HVAC zone energy while still providing thermal comfort and to reduce peak loads in small commercial buildings. We demonstrated cloud-based energy analytics, implemented a schedule and a Model Predictive Controller in a small commercial building to optimize HVAC energy, occupancy and electricity price. Initial integration of these technological innovations was performed through the creation of execution containers containing the WAVE agent and various driver, proxy, or building system function logic. The research added to the understanding of efficient sensor hardware, secure middleware, time-series data management (high performance database), efficient communication protocols, and interoperating with applications and building systems. The project showed the technical effectiveness and economic feasibility of creating a low-cost, modular, and easy-to-deploy sensor. Through conversations with multiple end users, the research team discovered that many customers wanted data management and services in addition to the sensors. HamiltonIOT developed packages of sensors, border router, and data services to provide a seamless “plug-and-play” sensor deployment. Some customers were willing to pay for higher quality sensors (such as light); some customers wanted a robust enclosure (waterproof).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

Enabling Real-Time Communication in Multi-Agent Systems: A Graph Neural Network Based Approach

Global connectivity enables effective coordination in Multi-Agent Systems (MAS). Solving these connection problems under hardware constraints is an NP-hard non-Euclidean Degree Constrained Minimum Spanning Tree (DCMST) problem. Prior MAS controllers coordinate team movement for task completion and collision avoidance; some considering Line-of-Sight (LOS) maintenance but prioritizing flexibility over guarantees. Evolutionary Algorithms (EA) have been shown to find good solutions for DCMST, but their performance degrades with larger populations required to support a large MAS. We present a method based on edge graph attention networks, trained offline to reduce online computation times. Empirical comparisons with greedy polynomial-time solvers and EA show that our method leverages latent graph information to consistently find constraint-satisfying solutions in less time.

connectivity maintenance↗
DOE OSTI2026

Self-Healing Lithium Dendrites through Spontaneous Passivating Layer Formation for Stable Solid-State Lithium–Metal Batteries

All-solid-state lithium–metal batteries have attracted significant attention, owing to their high energy density and superior safety. However, lithium–metal penetration through the solid electrolyte, leading to short-circuiting, remains a critical failure mode that demands comprehensive mitigation strategies. Most existing strategies are effective only prior to the initiation of lithium-dendrite formation and fail once dendrites begin to propagate through the electrolyte. In this study, we propose a self-healing mechanism in which the penetrated lithium reacts with a self-healing agent to form a passivating layer along the particle boundaries. Lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) was incorporated into a Li 6 PS 5 Cl solid electrolyte as the self-healing agent to suppress lithium-dendrite propagation even after dendrite formation initiated under high current densities. The self-healing induced by LiTFSI was verified through comprehensive experimental analyses and was further demonstrated in a full-cell configuration. Moreover, LiTFSI incorporation plays an important role in increasing the critical current density by reducing the overall electronic conductivity of the solid electrolyte and facilitating the formation of a robust LiF-containing solid-electrolyte interphase.

25 ENERGY STORAGE↗
DOE OSTI2025

A review on bacteria-derived antioxidant metabolites: their production, purification, characterization, potential applications, and limitations

Abstract Antioxidants are organic molecules that scavenge reactive oxygen species (ROS) and reactive nitrogen species (RNS), thereby maintaining cellular redox balance in living organisms. The human body synthesizes endogenous antioxidants, whereas humans obtain exogenous antioxidants from other organisms such as plants, animals, fungi, and bacteria. This review primarily focuses on the antioxidant potential of natural metabolites and extracts from five major bacterial phyla, including the well-studiedActinobacteriaandCyanobacteria, as well as less-studiedBacteroides,Firmicutes, andProteobacteria.The literature survey revealed that the metabolites and the extracts with antioxidant activity can be obtained from bacterial cells and their culture supernatants. The metabolites with antioxidant activity include pigments, phycobiliproteins, polysaccharides, mycosporins-like amino acids, peptides, phenolic compounds, and alkaloids. Both metabolites and extracts demonstrate in vitro antioxidant capacity through radical-scavenging, metal-reducing, and metal-chelating activity assays. In in vivo models, they can scavenge ROS and RNS directly and/or indirectly eliminate them by enhancing the activities of antioxidant enzymes, such as catalase, superoxide dismutase, and glutathione peroxidase. Due to their antioxidant activities, they may find applications in the cosmetic industry as anti-aging agents for the skin and in medicine as drugs or supplements for combating oxidative stress-related disorders, such as neurodegenerative diseases and diabetes. The literature survey also elucidated that some metabolites and extracts with antioxidant activity also exhibited strong antimicrobial properties. Therefore, we consider that they may have future applications in the treatment of infectious diseases, the preparation of pathogen-free healthy foods, and the extension of food shelf life.

Pharmacology & Pharmacy↗
DOE OSTI2020

Rational Polymer Design of Stretchable Poly(ionic liquid) Membranes for Dual Applications

Many functional polymeric materials are inherently fragile at ambient conditions. Consequently, making them elastic and flexible is a challenging task, and such achievement is especially meaningful in a wide range of applications including separation membranes and stretchable devices. Poly(ionic liquids) (PILs), such as vinyl-imidazolium-based polymers, are known to be “brittle” functional polymers due to their “glassy” nature at ambient temperature. We herein developed a viable approach to enable glassy PILs with high flexibility and good elasticity via a rational molecular design of chemical composition and polymer architectures. The reversible addition/fragmentation chain transfer agents (RAFT-CTAs) were attached to the flexible poly(dimethylsiloxane) (PDMS) backbones. The polymerization of functional ionic liquid monomers from RAFT-CTAs provided grafted copolymers with the functional side chains, which were further cross-linked by di-functional PDMS. Poly(ethylene glycol) methacrylate is copolymerized with ionic liquid monomers to reduce the glass transition temperature (Tg), providing higher chain mobility and elasticity at ambient temperature. The synthesized elastic PIL-based membranes (E-PILs) have dramatically improved stretchability, reaching 122–422%. In addition to significantly improved extensibility, the synthesized E-PILs also exhibit higher ionic conductivity, critical for potential applications in solid-state batteries. Moreover, in comparison to glassy solid PIL membranes, the E-PILs also exhibited enhanced flexibility and excellent gas-separation performance.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Advanced spectrometric methods for characterizing bio-oils to enable refineries to reduce fuel carbon intensity during co-processing

A promising approach for supplementing petroleum-derived fuels to support reductions in green-house gas emissions is to convert abundant biomass feedstocks into renewable carbon-rich oils using pyrolysis. However, the resultant bio-oils contain various oxygenated compounds that can impart acidity, chemical and thermal instability, and immiscibility with petroleum derived fuels, necessitating further upgrading to derive fuel blendstocks. Co-processing bio-oils and petroleum-derived liquids in existing refineries is a potentially near-term, cost-effective approach for upgrading bio-oils while reducing refinery carbon intensities. However, one cause for hesitation in co-processing bio-oils is limited comprehensive characterization and speciation of the bio-oil components. Advanced analytical techniques are currently under investigation to enable identification of elusive species in bio-oils, enabling researchers to develop strategies to mitigate catalyst deactivation agents and contaminants. This review provides a brief overview of several analytical methods commonly used to analyze bio-oils and their limitations. In addition, advanced techniques currently under development are discussed to further elucidate bio-oil components that may limit its end use. This will help inform the technical and economic feasibility of co-processing bio-oils with petroleum-derived liquids, therefore, improving the overall downstream processes for biofuels blendstock production.

09 BIOMASS FUELS↗
DOE OSTI2024

E. coli -expressed SECRET AGENT O -GlcNAc modifies threonine 829 of GIGANTEA

The Arabidopsis thaliana glycosyl transferases SPINDLY (SPY) and SECRET AGENT (SEC) modify nuclear and cytosolic proteins with O-linked fucose or O-linked N-acetylglucosamine (O-GlcNAc), respectively. O-fucose and O-GlcNAc modifications can occur at the same sites. SPY interacts physically and genetically with GIGANTEA (GI), suggesting that it could be modified by both enzymes. Previously, we found that, when co-expressed in Escherichia coli, SEC modifies GI; however, the modification site was not determined. By analyzing the overlapping sub-fragments of GI, we identified a region that was modified by SEC in E. coli. Modification was undetectable when threonine 829 (T829) was mutated to alanine, while the T834A and T837A mutations reduced the modification, suggesting that T829 was the primary or the only modification site. Mapping using mass spectrometry detected only the modification of T829. Previous studies have shown that the positions modified by SEC in E. coli are modified in planta, suggesting that T829 is O-GlcNAc modified in planta.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Nested active learning for efficient model contextualization and parameterization: pathway to generating simulated populations using multi-scale computational models

There is increasing interest in the use of mechanism-based multi-scale computational models (such as agent-based models (ABMs)) to generate simulated clinical populations in order to discover and evaluate potential diagnostic and therapeutic modalities. The description of the environment in which a biomedical simulation operates (model context) and parameterization of internal model rules (model content) requires the optimization of a large number of free parameters. In this work, we utilize a nested active learning (AL) workflow to efficiently parameterize and contextualize an ABM of systemic inflammation used to examine sepsis. Contextual parameter space was examined using four parameters external to the model’s rule set. The model’s internal parameterization, which represents gene expression and associated cellular behaviors, was explored through the augmentation or inhibition of signaling pathways for 12 signaling mediators associated with inflammation and wound healing. We have implemented a nested AL approach in which the clinically relevant (CR) model environment space for a given internal model parameterization is mapped using a small Artificial Neural Network (ANN). The outer AL level workflow is a larger ANN that uses AL to efficiently regress the volume and centroid location of the CR space given by a single internal parameterization. We have reduced the number of simulations required to efficiently map the CR parameter space of this model by approximately 99%. In addition, we have shown that more complex models with a larger number of variables may expect further improvements in efficiency.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Discovery and characterization of a selective IKZF2 glue degrader for cancer immunotherapy

Malignant tumors can evade destruction by the immune system by attracting immune-suppressive regulatory T cells (Treg) cells. The IKZF2 (Helios) transcription factor plays a crucial role in maintaining function and stability of Treg cells, and IKZF2 deficiency reduces tumor growth in mice. Here we report the discovery of NVP-DKY709, a selective molecular glue degrader of IKZF2 that spares IKZF1/3. We describe the recruitment-guided medicinal chemistry campaign leading to NVP-DKY709 that redirected the degradation selectivity of cereblon (CRBN) binders from IKZF1 toward IKZF2. Selectivity of NVP-DKY709 for IKZF2 was rationalized by analyzing the DDB1:CRBN:NVP-DKY709:IKZF2(ZF2 or ZF2-3) ternary complex X-ray structures. Exposure to NVP-DKY709 reduced the suppressive activity of human Treg cells and rescued cytokine production in exhausted T-effector cells. In vivo, treatment with NVP-DKY709 delayed tumor growth in mice with a humanized immune system and enhanced immunization responses in cynomolgus monkeys. NVP-DKY709 is being investigated in the clinic as an immune-enhancing agent for cancer immunotherapy.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Hydrotrope-enabled high concentration aqueous electrolytes for reversible and sustainable iron metal anodes

Iron metal-based energy storage devices hold great potential in stationary grid-scale sustainable energy due to the high theoretical specific capacity, ultralow cost, and abundance of iron. However, their practical deployment is limited by the poor reversibility of iron plating and stripping, as well as competitive hydrogen evolution. Here we introduce the concept of hydrotropy into iron electrolytes by developing an environmentally friendly and cost-effective high-concentration ferrous sulfate electrolyte using urea as a hydrotropic agent. The designed electrolyte increases the Coulombic efficiency of iron metal electrodes to approximately 96.5%, compared with ~84.6% for the dilute electrolyte. Molecular dynamics simulations and Raman spectroscopy illustrate that urea regulates the competitive coordination of anions and urea in the iron solvation sheath, while reconstructing the hydrogen-bond network in free water molecules. This reduces the activity of both solvated and free water, thereby alleviating hydrogen evolution. Moreover, the coordinated anions and urea molecules facilitate the in-situ formation of an organic-inorganic hybrid protective layer on the metallic iron, establishing a physical barrier against water and promoting homogeneous interfacial reactions. This work demonstrates an appealing opportunity to design cost-effective and high-performance electrolytes and propels the practical application of iron metal-based energy storage devices.

Feng, Guangxia [Stanford Univ., CA (United States)↗
DOE OSTI2026

Data-driven Mori–Zwanzig modeling of Lagrangian particle dynamics in turbulent flows

The dynamics of Lagrangian particles in turbulence play a crucial role in mixing, transport, and dispersion in complex flows. Their trajectories exhibit highly nontrivial statistical behavior, motivating the development of surrogate models that can reproduce these trajectories without incurring the high computational cost of direct numerical simulations of the full Eulerian field. This task is particularly challenging because reduced-order models typically lack access to the full set of interactions with the underlying turbulent field. Novel data-driven machine learning techniques can be powerful in capturing and reproducing complex statistics of the reduced-order/surrogate dynamics. In this work, we show how one can learn a surrogate dynamical system that is able to evolve a turbulent Lagrangian trajectory in a way that is point-wise accurate for short-time predictions (with respect to Kolmogorov time) and stable and statistically accurate at long times. This approach is based on the Mori–Zwanzig formalism, which prescribes a mathematical decomposition of the full dynamical system into resolved dynamics that depend on the current state and the past history of a reduced set of observables, and the unresolved orthogonal dynamics due to unresolved degrees of freedom of the initial state. We show how by training this reduced order model on a point-wise error metric on short time-prediction, we are able to correctly learn the dynamics of Lagrangian turbulence, such that also the long-time statistical behavior is stably recovered at test time. This opens up a range of applications, for example, for the control of active Lagrangian agents in turbulence.

97 MATHEMATICS AND COMPUTING↗