Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “DECISION THEORY”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Hierarchical Multi-agent Large Language Model Reasoning for Autonomous Heterogeneous Catalyst Discovery

Artificial intelligence is reshaping scientific exploration, but most methods automate procedural tasks without engaging in scientific reasoning, limiting autonomy in discovery. We demonstrate that hierarchical agentic large language model reasoning can efficiently drive simulation and scientific exploration. Across two chemical applications, CO adsorption on Cu surface transition metal adatoms and on M–N–C catalysts, reasoning-guided exploration reduces required atomistic simulations by up to 90% relative to heuristic or random selection. Comparisons across single-agent, multi-agent, and stochastic baselines show that hierarchical strategies yield more coherent and information-efficient search trajectories. Reasoning traces reveal chemically grounded decisions that cannot be explained by semantic bias or stochastic sampling. We realize these agentic reasoning strategies in Materials Agents for Simulation and Theory in Electronic-structure Reasoning (MASTER), a multimodal system that translates natural language into density functional theory workflows. Altogether, multi-agent collaboration accelerates heterogeneous catalyst discovery and marks a step toward more autonomous, reasoning-guided scientific exploration.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2025

Integrating Analytical Solutions and U-Net Model for Predicting Groundwater Contaminant Plumes in Pump-and-Treat Systems

Pump-and-treat (P&T) is a common technique for groundwater remediation involving the extraction and treatment of contaminated water above ground. Optimizing the design and operation of the P&T well network is essential for maximizing the system’s effectiveness and efficiency. However, this optimization often necessitates many model evaluations, leading to computationally demanding tasks. This study introduces a novel approach that integrates analytical solutions for groundwater dynamics with the U-Net (Ronneberger et al., 2015) deep learning framework to predict groundwater contaminant plume migration under dynamic pumping conditions. By incorporating the Thiem equation (Thiem, 1906) into the input preprocessing, the U-Net model transforms sparse well data into a continuous spatial field that captures the hydraulic impacts of pumping activities. This integration enables the model to leverage both deep learning capabilities and classical physics-based groundwater theories, enhancing prediction accuracy and computational efficiency. These advancements can facilitate rapid, large-scale evaluations of P&T optimization simulations, allowing for timely and effective decision-making in well placement and system management. We demonstrate the model's robust performance across both simplified transient 2D models and a more complex 3D heterogeneous site model at the 200 West P&T facility at the Hanford Site. The U-Net-based model offers substantial computational advantages, reducing simulation times significantly compared to full physics-based models and providing a powerful tool for rapid site evaluation and P&T system optimization, such as evaluating alternative P&T well network designs. Our findings highlight the potential of advanced machine learning models to significantly enhance the efficiency and sustainability of groundwater remediation efforts, offering a novel application of U-Net architecture in environmental science.

Pump-and-treat↗
DOE OSTI2025

Multi-scale, Multi-disciplinary, and Multi-agent Explainable AI with Koopman-Undergirded Learning, Prediction, and Analysis (M3EA KULPA) (Project Closeout Report)

The goal of this project was to develop and use domain-aware machine learning formulations, based on the Koopman Operator (KO), for modelling multi-scale, multi-disciplinary (e.g., multi-physics), and/or multi-agent systems. The project developed these formulations for the following cases: • Systems with dynamics at two separate time scales, • Systems with a bi-level hierarchical control structure, • Systems with bi-level hierarchical control and dynamics at two separate time scales (the lower level controls operating at the faster time scale), and • Systems with n separate but interacting agents/disciplines (with/without control, respectively); the controls for each agent could include bi-level hierarchical control and dynamics at two separate time scales as described above. The project then defined a set of dynamical systems consisting of different nonlinear oscillators that could be used to test these different formulations and then subsequently learned the KO models for those systems. With the KO models, we were able to do the following: • Quantify system stability, including both long-term and transient behavior, • Quantify the effects of feedbacks between the different time scales and agents/disciplines in terms of those feedbacks’ effects on system stability, • Replace a standard Proportional-Integral (PI) control in the hierarchical control structure with a KO-based Linear-Quadratic Regular (LQR), a form of optimal control, • Calculate optimal supervisory control policies a) with and without time scale separated dynamics at the lower level control levels and b) with both PI and KO-based LQR lower level control policies, and • Calculate dynamic Nash equilibria for multi-agent systems where each agent makes its own control decisions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Game Theory Approaches for System-level Incentive Design

This report presents a generalized Stackelberg game framework for designing and evaluating financial incentives that enhance power system resilience through strategic deployment of distributed energy resources(DERs) under various contingencies. The proposed approach addresses the challenge of coordinating individual community investment decisions to meet system-wide resilience objectives. The framework is demonstrated in a three-community test system subjected to two transmission contingency scenarios: inter-community line failure (Case 1) and complete main grid disconnection (Case 2). In both cases, three incentive levels are compared: a Base case with no financial incentives, and low and high incentive cases. In Case 1, the Base case (no incentives) results in a total installed DER capacity of 217.2 MW, with no load shedding due to alternative routing, but community costs remain high. Increasing incentives raises DER deployment to 286.9 MW, lowers aggregate community costs by $22M annually, and completely avoids the need for costly new transmission line construction. In Case 2, the Base case results in 24.3 MWh of unserved load; introducing incentives eliminates all load shedding and ensures up to 89 MWh of battery storage is available for emergency reserve. These results demonstrate that targeted incentives can dramatically improve grid resilience and cost-effectiveness. The framework thus offers policymakers and system planners a robust tool to quantify and compare the effectiveness of incentive programs for multi-community transmission networks behavior, system resilience, and economic efficiency.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Phonon Olympics: Phonon property and lattice thermal conductivity benchmarking from open-source packages

Three widely used open-source packages for determining phonon properties and lattice thermal conductivities (ALAMODE, phono3py, and ShengBTE) are benchmarked by teams of expert users and the package developers. The phonons for Ge, RbBr, monolayer MoSe 2 , and AlN are modeled at zero temperature, and they scatter through three-phonon and phonon-isotope processes, with thermal conductivities obtained from the linearized Peierls–Boltzmann transport equation with input from density functional theory calculations. Over a wide range of temperatures, the thermal conductivities calculated by the teams fall within at most ±15% of their mean values for each of the four materials. The phonon frequencies, obtained from the harmonic force constants, do not show large differences between the calculations, indicating that the modal heat capacities and group velocities are not responsible for the thermal conductivity variations. It is the lifetimes associated with three-phonon scattering, obtained from the cubic force constants, that drive the variations. The many decisions required to calculate the cubic force constants (e.g., supercell size, atomic displacement, neighbor cutoff, and application of symmetries) make identification of the precise origin of the thermal conductivity variations challenging. The calculated thermal conductivities do not generally show agreement with experimental measurements, which is attributed to the limitations of the density functional theory calculations. Guidance for the development of best practices is provided, which will help to standardize protocols needed for building thermal conductivity databases. The results provide a baseline for future benchmarking of other packages and more advanced calculations.

McGaughey, Alan J. H. [Carnegie Mellon Univ., Pitt↗
DOE OSTI2024

Optimal experimental design: Formulations and computations

Questions of ‘how best to acquire data’ are essential to modelling and prediction in the natural and social sciences, engineering applications, and beyond. Optimal experimental design (OED) formalizes these questions and creates computational methods to answer them. This article presents a systematic survey of modern OED, from its foundations in classical design theory to current research involving OED for complex models. We begin by reviewing criteria used to formulate an OED problem and thus to encode the goal of performing an experiment. We emphasize the flexibility of the Bayesian and decision-theoretic approach, which encompasses information-based criteria that are well-suited to nonlinear and non-Gaussian statistical models. We then discuss methods for estimating or bounding the values of these design criteria; this endeavour can be quite challenging due to strong nonlinearities, high parameter dimension, large per-sample costs, or settings where the model is implicit. A complementary set of computational issues involves optimization methods used to find a design; we discuss such methods in the discrete (combinatorial) setting of observation selection and in settings where an exact design can be continuously parametrized. Finally we present emerging methods for sequential OED that build non-myopic design policies, rather than explicit designs; these methods naturally adapt to the outcomes of past experiments in proposing new experiments, while seeking coordination among all experiments to be performed. Throughout, we highlight important open questions and challenges.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Search for charged-lepton flavor violation in the production and decay of top quarks using trilepton final states in proton-proton collisions at $\sqrt{s}$ =13 TeV

A search is performed for charged-lepton flavor violating processes in top quark (𝑡) production and decay. The data were collected by the CMS experiment from proton-proton collisions at a center-of-mass energy of 13 TeV and correspond to an integrated luminosity of 138 fb −1 . The selected events are required to contain one opposite-sign electron-muon pair, a third charged lepton (electron or muon), and at least one jet of which no more than one is associated with a bottom quark. Boosted decision trees are used to distinguish signal from background, exploiting differences in the kinematics of the final states particles. The data are consistent with the standard model expectation. Upper limits at 95% confidence level are placed in the context of effective field theory on the Wilson coefficients, which range between 0.024–0.424 TeV −2 depending on the flavor of the associated light quark and the Lorentz structure of the interaction. These limits are converted to upper limits on branching fractions involving up (charm) quarks, 𝑡 → 𝑒⁢𝜇⁢𝑢 (𝑡 → 𝑒⁢𝜇⁢𝑐), of 0.032⁢(0.498) × 10 −6 , 0.022⁢(0.369) × 10 −6 , and 0.012⁢(0.216) × 10 −6 for tensorlike, vectorlike, and scalarlike interactions, respectively.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

TeV-scale lepton number violation: Connecting leptogenesis, neutrinoless double beta decay, and colliders

In the context of TeV-scale lepton number violating (LNV) interactions, we illustrate the interplay between leptogenesis, neutrinoless double beta ( 0 ν β β ) decay, and LNV searches at proton-proton colliders. Using a concrete model for illustration, we overcome the limitations of previous effective field theory analyses and are able to identify the parameter space where standard thermal leptogenesis is rendered unviable due to washout processes. Moreover, we show how 0 ν β β decay and p p collisions provide complementary probes. We find that the new particle spectrum can have a decisive impact on the relative sensitivity of these two probes. Published by the American Physical Society 2024

Astronomy & Astrophysics↗
DOE OSTI2025

HydroForecast Long-term: Improving hydropower’s resilience to climate change through accurate climate-scale

With hydrologic patterns and water availability across the globe shifting due to climate change, advancements in hydrologic prediction systems can help significantly reduce the uncertainties that utilities and water supply entities have in their decision making. Understanding and estimating hydrology at the climate scale is critical for managing water resources under changing climate scenarios. This project focuses on integrating state-of-the-art neural network modeling with downscaled climate projections to deliver the reliable water supply projections decades into the future to meet an urgent need from hydropower operators and water utilities. In this Phase 1 DOE SBIR proposal, we developed and validated a theory-guided neural network model, HydroForecast Long-term, for climate-scale hydrology and implemented the model within existing HydroForecast infrastructure. HydroForecast Long-term combines the most accurate streamflow modeling system with a flexible and scalable data architecture to generate water supply projections out to the year 2100. This report illustrates that we have achieved our four objectives: 1) create a prototype of HydroForecast Long-term, building the neural network prediction model, 2) build an automated data input pipeline that processes large amounts of data from the latest global temperature and precipitation climate models; 3) benchmark the accuracy of the hydrologic model over the recent two decades over a large set of diverse basins, and 4) create a set of output visuals and summary metrics informed by customer feedback that connect the data to critical decision points. This work empowers water users to make data-informed decisions supporting a resilient, renewable-powered grid and water system. The results advance the Department of Energy’s mission by addressing critical gaps in water supply planning under climate change.

13 HYDRO ENERGY↗
DOE OSTI2024

Center For Actinide Science and Technology (CAST)

CAST encompassed 17 senior researchers in its first funding cycle and added two additional theorists briefly in the final two years. Approximately 37 graduate students and post-doctoral associates were supported by CAST. Within the 19 senior researchers there were changes in composition since the center was first funded. One researcher left associated with a move to another university. Two were phased out because they did not meet the expectations of CAST in terms of research performance. These decisions were always made in consultation with our External Advisory Committee (EAC) and DOE management. In fact, the phasing out of the two PI’s was suggested first by the EAC. Two new researchers were also brought in response to feedback at our midterm review where it was suggested that we needed increased theory support in CAST. No further changes were implemented since then, and CAST researchers thought that stability and optimal productivity was achieved.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

NANO.PTML model for read-across prediction of nanosystems in neurosciences. computational model and experimental case of study

Abstract Neurodegenerative diseases involve progressive neuronal death. Traditional treatments often struggle due to solubility, bioavailability, and crossing the Blood-Brain Barrier (BBB). Nanoparticles (NPs) in biomedical field are garnering growing attention as neurodegenerative disease drugs (NDDs) carrier to the central nervous system. Here, we introduced computational and experimental analysis. In the computational study, a specific IFPTML technique was used, which combined Information Fusion (IF) + Perturbation Theory (PT) + Machine Learning (ML) to select the most promising Nanoparticle Neuronal Disease Drug Delivery (N2D3) systems. For the application of IFPTML model in the nanoscience, NANO.PTML is used. IF-process was carried out between 4403 NDDs assays and 260 cytotoxicity NP assays conducting a dataset of 500,000 cases. The optimal IFPTML was the Decision Tree (DT) algorithm which shown satisfactory performance with specificity values of 96.4% and 96.2%, and sensitivity values of 79.3% and 75.7% in the training (375k/75%) and validation (125k/25%) set. Moreover, the DT model obtained Area Under Receiver Operating Characteristic (AUROC) scores of 0.97 and 0.96 in the training and validation series, highlighting its effectiveness in classification tasks. In the experimental part, two samples of NPs (Fe 3 O 4 _A and Fe 3 O 4 _B) were synthesized by thermal decomposition of an iron(III) oleate (FeOl) precursor and structurally characterized by different methods. Additionally, in order to make the as-synthesized hydrophobic NPs (Fe 3 O 4 _A and Fe 3 O 4 _B) soluble in water the amphiphilic CTAB (Cetyl Trimethyl Ammonium Bromide) molecule was employed. Therefore, to conduct a study with a wider range of NP system variants, an experimental illustrative simulation experiment was performed using the IFPTML-DT model. For this, a set of 500,000 prediction dataset was created. The outcome of this experiment highlighted certain NANO.PTML systems as promising candidates for further investigation. The NANO.PTML approach holds potential to accelerate experimental investigations and offer initial insights into various NP and NDDs compounds, serving as an efficient alternative to time-consuming trial-and-error procedures.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Multidimensional perspectives of geo-epidemiology: from interdisciplinary learning and research to cost–benefit oriented decision-making

Research typically promotes two types of outcomes (inventions and discoveries), which induce a virtuous cycle: something suspected or desired (not previously demonstrated) may become known or feasible once a new tool or procedure is invented and, later, the use of this invention may discover new knowledge. Research also promotes the opposite sequence—from new knowledge to new inventions. This bidirectional process is observed in geo-referenced epidemiology—a field that relates to but may also differ from spatial epidemiology. Geo-epidemiology encompasses several theories and technologies that promote inter/transdisciplinary knowledge integration, education, and research in population health. Based on visual examples derived from geo-referenced studies on epidemics and epizootics, this report demonstrates that this field may extract more (geographically related) information than simple spatial analyses, which then supports more effective and/or less costly interventions. Actual (not simulated) bio-geo-temporal interactions (never captured before the emergence of technologies that analyze geo-referenced data, such as geographical information systems) can now address research questions that relate to several fields, such as Network Theory. Thus, a new opportunity arises before us, which exceeds research: it also demands knowledge integration across disciplines as well as novel educational programs which, to be biomedically and socially justified, should demonstrate cost-effectiveness. Grounded on many bio-temporal-georeferenced examples, this report reviews the literature that supports this hypothesis: novel educational programs that focus on geo-referenced epidemic data may help generate cost-effective policies that prevent or control disease dissemination.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Nonadiabatic dynamics of photoexcited thiopyridone isomers: An interplay between El-Sayed’s conditions and energy gap law

We investigated the non-adiabatic dynamics of photoexcited thiopyridone systems across their ortho-, meta-, and para-isomeric forms. The relaxation pathways of the three isomers in both gas phase and solvent environments are mapped using surface hopping dynamics based on time-dependent density functional theory. Our analysis highlights the influence of isomeric structures on photophysical behavior, offering insights into design principles to control photochemical phenomena. The simulations suggest a systematic reduction in the rate of intersystem crossing (ISC) from ortho- to meta- to para-isomer. Comparisons with multiconfigurational wave function methods in the gas phase further demonstrate how electronic structure influences the predicted dynamical pathways. The simulated dynamics demonstrates that the spin–orbit coupling strength alone does not determine the rate of ISC, as both state energetics and underlying electronic and structural features play decisive roles. These aspects explain the much slower ISC in the para-isomer, as well as the non-negligible role of the El-Sayed forbidden pathway in the computed ISC dynamics.

Complete-active space self-consistent field↗
DOE OSTI2024

Promoting regulatory acceptance of combined ion and neutron irradiation testing of nuclear reactor materials: Modeling and software considerations

As the needs for the nuclear energy industry continue to evolve in the 21st century, timely adoption of new technological solutions acceptable to regulatory agencies is critical. Quantitative prediction of radiation damage in materials and its impact on mechanical properties is a key component of licensing and regulatory decisions regarding nuclear power plants. Accelerated testing methodologies such as combined ion and neutron irradiation data sets are crucial for the development and deployment of new materials and new manufacturing methods (e.g., additive manufacturing). However, regulatory acceptance of accelerated testing methodologies is necessary for their adoption. Further, the present work discusses the fundamental basis for comparing ion- and neutron-induced material microstructures, the theory behind interpreting radiation damage across length and time scales and radiation types, and the codes, standards, and quality assurance concerns surrounding different modeling methods and software. In particular, recommendations are given as to the path forward that will enable national laboratories, academia, and industry to develop the modeling and software basis for regulatory acceptance of the combined use of ion and neutron irradiation for material performance evaluation.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2024

Description of FY25 Theory and Simulation Performance Target: Development of an integrated modeling framework for fusion reactor design and assessment

The urgency to deliver fusion power is growing now more than ever, with increasing pressure for both public programs and private companies to meet milestones timelines and overcome significant remaining technical challenges to ensure growth of a nascent fusion industry in time to meet rapidly growing clean energy demands. With incredible advancements in computation and years of investment in fusion model development and validation, integrated modeling is poised to fill a key role in accelerating the timeline to a fusion pilot plant (FPP). Future fusion pilot plants will operate in regimes far beyond current experience, and device design will rely on physics-based prediction and extrapolation. Many concepts will also rely on simulation to assess safety (shielding, tritium management, materials activation and lifetimes), economics and scalability before the decision to build. Importantly, integrated simulation can be used to reveal and solve the complexities of system integration that may otherwise not be apparent in physical components or models developed in isolation. New experimental test facilities that produce relevant conditions to validate and resolve key technical challenges for various subsystems (materials, blankets, fuel cycle, etc.) have been repeatedly called for by the fusion community but are not yet realized. Integrated modeling has an important role in identifying realistic load conditions (thermal, electromagnetic, plasma, neutron and photon loads, etc.) and defining the components and experiments for these test facilities in order to ensure meaningful validation that sufficiently reduces modeling uncertainties and technical risk for the full integrated reactor. The Fusion REactor Design and Assessment (FREDA) SciDAC project is building a component-based integrated modeling framework & data structure to enable self-consistent, multi-fidelity, iterative optimization workflows for the fusion reactor design process. FREDA aims to shorten the time to viable designs by providing a set of flexible workflows to support the various stages of the design process using an integrated model hierarchy, ranging from the simple analytic descriptions to the highest fidelity, theory-based plasma and engineering modeling developed by the fusion and fission communities. These tools are expected to be needed for timely support of FPP design in the milestone program and in the FIRE collaboratives. The plasma simulation backbone of FREDA is IPS-FASTRAN with newly developed coupled Core-Edge Pedestal-SOL (CESOL) workflows, which is being extended to the far-SOL region up to the plasma facing components. FREDA incorporates the FERMI engineering modeling suite and will enable self-consistent evaluation of the thermal shields, limiters, blanket, magnets, and other surrounding structures with predictions of temperatures, erosion, dpa, activation, tritium generation and transport, creep, corrosion, material degradation, etc. Parametric generation of 3D CAD enables rapid iteration of component geometry in response to plasma and loading specifications.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Toward equitable environmental exposure modeling through convergence of data, open, and citizen sciences: an example of air pollution exposure modeling amidst increasing wildfire smoke

Exposure modeling is critical in environmental epidemiology and human health but may face challenges (e.g., skewed data, unequal error, context-insensitive validation, and computational demands). Modeling decisions reflect the intended use of the models and the values that modelers prioritize. We aimed to provide a conceptual framework and machine learning (ML) modeling protocols that address these issues. With 500m-gridded hourly PM 2.5 and O 3 levels in Illinois before, during, and after the 2023 Canadian wildfire season as a motivating example, we conducted modeling experiments to evaluate modeling methods, guided by three domains we propose based on theories of science: 1) Data Diversity, leveraging open and citizen science data to enhance inclusivity, parsimony, and representativeness; 2) Equitable Accuracy, ensuring fairly distributed uncertainties across subpopulations; and 3) Sustainable Modeling, balancing accuracy with reducing computational demands to promote accessibility for under-resourced researchers. Here, we found that ML with publicly available data can achieve high accuracy. Depending on methods, performance may vary substantially, even with identical input data. Large but skewed data may reduce performance. Misuse of cross-validation protocols can underestimate prediction error; although we observed R 2 s of ∼98 %, the modeled estimates varied significantly, indicating the need for careful model validation. By using new modeling protocols including representativeness-considered training and validation data and a new loss function, we achieved high agreement between estimates and ground-based measurements (e.g., R 2 = ∼90 % for PM 2.5 ; ∼80 % for O 3 ), equally distributed errors across sociodemographic strata and urban–rural divides, and reduction in computation time—from several weeks or months to a few days.

Exposure assessment↗
DOE OSTI2024

Single-shot Quantum Signal Processing Interferometry

Quantum systems of infinite dimension, such as bosonic oscillators, provide vast resources for quantum sensing. Yet, a general theory on how to manipulate such bosonic modes for sensing beyond parameter estimation is unknown. We present a general algorithmic framework, quantum signal processing interferometry (QSPI), for quantum sensing at the fundamental limits of quantum mechanics by generalizing Ramsey-type interferometry. Our QSPI sensing protocol relies on performing nonlinear polynomial transformations on the oscillator's quadrature operators by generalizing quantum signal processing (QSP) from qubits to hybrid qubit-oscillator systems. We use our QSPI sensing framework to make efficient binary decisions on a displacement channel in the single-shot limit. Theoretical analysis suggests the sensing accuracy, given a single-shot qubit measurement, scales inversely with the sensing time or circuit depth of the algorithm. We further concatenate a series of such binary decisions to perform parameter estimation in a bit-by-bit fashion. Numerical simulations are performed to support these statements. Our QSPI protocol offers a unified framework for quantum sensing using continuous-variable bosonic systems beyond parameter estimation and establishes a promising avenue toward efficient and scalable quantum control and quantum sensing schemes beyond the NISQ era.

Physics↗
DOE OSTI2024

Metric Learning to Accelerate Convergence of Operator Splitting Methods

Recent developments in machine learning have led to promising advances in accelerating the solution of constrained optimization problems. Increasing demand for real-time decision-making capabilities in applications such as artificial intelligence and optimal control has led to a variety of proposed strategies for learning to produce fast solutions to optimization problems. For example, recent works have shown that it is possible to accelerate the convergence of optimization algorithms by learning to select their parameters, such as gradient descent stepsizes. This work proposes a new approach, in which the underlying metric spaces of proximal operator splitting algorithms are learned to maximize convergence rate. While prior works in optimization theory have derived optimal metrics in simple cases, no such result exists for many practical problem forms including general Quadratic Programming (QP). This paper shows how differentiable optimization can enable the end-to-end learning of proximal metrics, enhancing the convergence of proximal algorithms for QP problems beyond what is possible based on known theory. Additionally, the results illustrate a strong connection between the learned proximal metrics and active constraints at the optima, leading to an interpretation in which the predicted proximal metrics can be viewed as a form of active set prediction.

King, Ethan [BATTELLE (PACIFIC NW LAB)]↗