Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Bayesian Evaluation Framework”

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 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Next-Level Energy Management in Manufacturing: Facility-Level Energy Digital Twin Framework Based on Machine Learning and Automated Data Collection

This research introduces an energy prediction framework at the facility level supported by automated data collection and machine learning models. It investigates whether reducing the prediction time scale allows for applying more complex machine learning techniques and if those techniques improve the prediction accuracy. The primary advantages of this framework lie in its automation of the energy prediction process and its provision of real-time energy data suitable for use in energy dashboards or digital twins. A sitewide dataset was created by combining 15 min energy and daily production data of five shops—assembly, battery, body (electric), body (gas), and paint—from a globally recognized electric vehicle manufacturer. Various machine learning models were evaluated on daily, weekly, and monthly datasets, including, in increasingly complex order: naïve, simple linear regression, net regularized generalized linear regression, principal component regression, k-nearest neighbor, random forest, and Bayesian regularized neural network. Compared to the current state-of-the-art energy consumption prediction for the industrial facility level, this research investigates more complex models and smaller time intervals for higher accuracy. The findings revealed that the more complex monthly models require a minimum of a year and a half of data to operate, while weekly models demand a year of data to achieve improved accuracy. Daily models can operate with only six months of data but exhibit poor performance due to reduced prediction accuracy of production. Key challenges identified include access to reliable, high-quality energy and production data and the initial demand for human labor.

digital twin↗
DOE OSTI2025

Analytical gradient-based optimization of CALPHAD model parameters

The calibration of CALPHAD (CALculation of PHAse Diagrams) models involves the solution of a very challenging high-dimensional multiobjective optimization problem. Traditional approaches to parameter fitting predominantly rely on gradient-free methods, which while robust, are computationally inefficient and often scale poorly with model complexity. In this work, we introduce and demonstrate a generalizable framework for analytic gradient-based optimization of the parameters of the CALPHAD model enabled by the recently formalized Jansson derivative technique. This method allows for efficient evaluation of gradients of thermodynamic properties at equilibrium with respect to model parameters, even in the presence of arbitrarily complex internal degrees of freedom. Leveraging these semi-analytic gradients, we employ the conjugate gradient (CG) method to optimize thermodynamic model parameters for four binary alloy systems: Cu-Mg, Fe-Ni, Cr-Ni, and Cr-Fe. Across all systems, CG achieves comparable or superior optimality relative to Bayesian ensemble Markov Chain Monte Carlo (MCMC) with improvements in computational efficiency ranging from one to three orders of magnitude. Furthermore, our results establish a new paradigm for CALPHAD assessments in which high fidelity data-rich model calibration becomes tractable using deterministic gradient-informed algorithms.

CALPHAD↗
NASA NTRS2019

Evaluating Biosignatures for Life Detection

The goal here is to develop a conceptual framework for evaluating the ability of different biosignatures to provide evidence for the presence of life in planned missions or observational studies. The focus is on intrinsic characteristics of biosignatures in a given space environment rather than on their detection, which depends on the current state of technology. Suitable evaluation procedures are based on an extensive body of previous work on related problems in which different options are evaluated to make decisions in business, engineering, medical fields and social or political arena. In these cases, three approaches have proven to be particularly useful. Two of them, signal detection theory (SDT) and Bayesian hypothesis testing are based on probabilities. The third approach, multi-attribute utility (MAU) is based on utility theory.

Pohorille, Andrew↗
DOE OSTI2023

Extending Parsimonious Bayesian Inference

Parsimonious Bayesian inference is a theoretical framework for efficient data assimilation that seeks to balance increased consistency between predictions and training data against corresponding increases in model complexity. Within this framework, over-training is understood as optimization that encodes excessive information within model parameters while only achieving small improvements between predictions and training data. This project aims to develop practical methods of limiting excess model information during optimization. One key observation is that practical heuristics for parsimonious learning in high-dimensions must balance expressivity, i.e. the ability of the model to capture diverse predictions with only a few non-zero parameters, against discoverability, i.e. the ability to train the model with gradient-based optimization and drive parameters to low information states. As such, we developed logical activation functions that are able to adaptively approximate arbitrary truth tables that define Boolean logic operations within a probabilistic framework. These functions have demonstrated the ability to learn exclusive disjunction (XOR) and conditioned disjunction (if [condition] then [result_if_true] else [result_if_false]) within a single layer of a neural network. To efficiently exploit these activation functions to drive parsimonious learning required several other advances within the domain of variational inference. The most efficient form of complexity suppression is structured sparsification, driving most model parameters to zero while achieving the structural coherence among nonzeros needed for bandwidth reduction. Such models are not only far more efficient at suppressing information-theoretic complexity, they also reduce the other forms of complexity (computations, communication, storage, and the number of dependencies needed to evaluate predictions). Aiming to support enhanced sparsification, this project examined new approaches to high-dimensional variational inference that allow us to calibrate and control parameter uncertainty during optimization. By identifying which parameters can sustain sparsifying perturbations with little impact on prediction quality, we can develop better pruning strategies by framing them as approximate Bayesian inference. These advances also open paths to mitigate concerns with deploying advanced learning methods in resource-constrained environments, such as running models on power-limited or communication-limited devices.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Filling the Gaps: A Bayesian Mixture Model for Imputing Missing Soil Water Content Data

ABSTRACT Soil water content (SWC) data are central to evaluating how soil moisture varies over time and space and influences critical plant and ecosystem functions, especially in water‐limited drylands. However, sensors that record SWC at high frequencies often malfunction, leading to incomplete timeseries and limiting our understanding of dryland ecosystem dynamics. We developed an analytical approach to impute missing SWC data, which we tested at six eddy flux tower sites along an elevation gradient in the southwestern United States. We impute missing data as a mixture of linearly interpolated SWC between the observed endpoints of a missing data gap and SWC simulated by an ecosystem water balance model (SOILWAT2). Within a Bayesian framework, we allowed the relative utility (mixture weight) of each component (linearly interpolated vs. SOILWAT2) to vary by depth, site and gap characteristics. We explored “fixed” weights versus “dynamic” weights that vary as a function of cumulative precipitation, average temperature, and time since the start of the gap. Both models estimated missing SWC data well ( R 2 = 0.70–0.88 vs. 0.75–0.91 for fixed vs. dynamic weights, respectively), but the utility of linearly interpolated versus SOILWAT2 values depended on site and depth. SOILWAT2 was more useful for more arid sites, shallower depths, longer and warmer gaps and gaps that received greater precipitation. Overall, the mixture model reliably gap‐fills SWC, while lending insight into processes governing SWC dynamics. This approach to impute missing data could be adapted to accommodate more than two mixture components and other types of environmental timeseries.

Ogle, Kiona [School of Informatics, Computing, and↗
NASA NTRS2026

Benchmarking Bayesian Optimization Frameworks and Acquisition Strategies for Materials Discovery and Autonomous Laboratories

Bayesian optimization (BO) can accelerate materials discovery by guiding expensive experiments toward the most promising processing conditions. We systematically compare five BO surrogate and framework combinations (Gaussian processes in Ax, Gaussian processes and Monte-Carlo neural networks in BayBE, random forests in Lolopy, and tree-structured Parzen (TPE) estimators in Hyperopt) on three benchmarks that mimic common materials design tasks (a discrete solid-electrolyte composition space, a hybrid discrete/continuous laminate-composite design problem solved with micromechanics modeling, and the continuous Ishigami analytic function which is a standard optimization benchmark). Each BO surrogate is paired with posterior mean, probability of improvement, and expected improvement acquisition functions and run for 100 trials from randomized initial samples with uniform random search providing a control. Across five random seeds per setting, BayBE’s Gaussian-process surrogate with expected improvement consistently reached ≥95 % of the known optimum in the fewest evaluations, while Lolopy’s random forest matched or exceeded GP performance on purely categorical or mixed spaces at a higher computational cost. Posterior mean alone often stagnated at local optima, underscoring the need for exploration, whereas probability and expected improvement balanced exploration and exploitation leading to better optimization in fewer trials. Execution times ranged from milliseconds for TPE to minutes for neural-network and random-forest surrogates. These results establish baseline expectations for BO in automated materials laboratories and highlight expected improvement with Gaussian processes as a reliable first choice, with random forests offering a strong alternative when categorical variables dominate. The benchmark suite and code are released to facilitate future surrogate, acquisition, and constraint-handling research in data-driven materials optimization.

Bayesian optimization↗
DOE OSTI2024

Bayesian Framework for Bioburden Density Estimation in Planetary Protection

To comply with the international planetary protection policy set forth by the Committee on Space Research and NASA Agency level requirements, spacecraft destined to biologically sensitive planetary bodies have to minimize terrestrial biological contamination. Analysis, testing and inspection are the standard forward verification activities that are used to demonstrate compliance with the biological contamination requirements. For testing of spacecraft surface areas, a swab or wipe sample is collected from surfaces prior to last access and subsequently processed in the lab using NASA Approved Planetary Protection Methods for Culture Based Assays. Raw data resulting from this assay is then statistically treated employing a mathematical paradigm stemming from the 1970’s Viking Lander Project to generate the bioburden density and total microbial bioburden present. This standard approach arbitrarily accounts for error and provides an upper conservative bound as it reports the maximum number of spores estimated to be present on flight hardware surfaces. A bioburden density estimate factors in the following variables: the observed bioburden count, representative volume processed, sampling efficiencies. Notably, to account for error in the approach, a 0 observed count is arbitrarily changed to a count of 1 for each hardware grouping. The data generated by spacecraft bioburden verification campaigns in the past have resulted in <80% of wipes and <90% of swabs containing a bioburden count of 0. As such, having a robust and well documented statistical approach for dealing with the probability of low incident rates is necessary to be able to estimate spacecraft bioburden. Being able to statistically describe the bioburden distribution and associated confidence level is a gamechanger for the development of bioburden allocations during mission design and will allow for tighter management of risk throughout spacecraft build. Thus, Empirical Bayes statistical approach was evaluated to estimate the microbial bioburden on spacecraft to mitigate the aforementioned mathematical concerns and provide a probabilistic bioburden distribution of the flight hardware surface. For application of this approach to performing bioburden calculations, a range of non-informative prior assumptions on hardware surfaces are explored for Bayesian analyses while informative priors using posterior distributions from prior assays are utilized for Empirical Bayes analyses. Several non-informative priors are currently under investigation to assess fitness including use of these priors to serve as a foundation to build off of NASA specification values or a basis of risk to account for unknowns during the integration and testing process. Informative priors under consideration are generated using sampled bioburden values from hardware originating within like processing environments (e.g. vendor cleaning process or similar assembly process), temporal spacecraft status events as a prediction for hardware cleanliness of future samples, and heritage system bioburden actuals to predict allocation for subsequent missions. Informative priors and probabilistic bioburden distributions are then validated using data sets from the Mars Exploration Rover, Mars Science Laboratory, and InSight missions. Using Empirical Bayes approach to generate a probabilistic bioburden distribution as demonstrated through mission use cases provides a valid approach for use in the end-to-end requirements verification process.

97 - MATHEMATICS AND COMPUTING↗
DOE OSTI2025

A framework to evaluate machine learning crystal stability predictions

The rapid adoption of machine learning in various scientific domains calls for the development of best practices and community agreed-upon benchmarking tasks and metrics. We present Matbench Discovery as an example evaluation framework for machine learning energy models, here applied as pre-filters to first-principles computed data in a high-throughput search for stable inorganic crystals. We address the disconnect between (1) thermodynamic stability and formation energy and (2) retrospective and prospective benchmarking for materials discovery. Alongside this paper, we publish a Python package to aid with future model submissions and a growing online leaderboard with adaptive user-defined weighting of various performance metrics allowing researchers to prioritize the metrics they value most. To answer the question of which machine learning methodology performs best at materials discovery, our initial release includes random forests, graph neural networks, one-shot predictors, iterative Bayesian optimizers and universal interatomic potentials. We highlight a misalignment between commonly used regression metrics and more task-relevant classification metrics for materials discovery. Accurate regressors are susceptible to unexpectedly high false-positive rates if those accurate predictions lie close to the decision boundary at 0 eV per atom above the convex hull. The benchmark results demonstrate that universal interatomic potentials have advanced sufficiently to effectively and cheaply pre-screen thermodynamic stable hypothetical materials in future expansions of high-throughput materials databases.

Riebesell, Janosh↗
DOE OSTI2025

Characterization of Coal Refuse Sites in West Virginia: Stream Loss, Volume Estimation, and Land Cover Analysis for Restoration Feasibility

Coal refuse disposal across Appalachia has resulted in widespread environmental degradation, including stream burial, landscape destabilization, and long-term hydrologic impacts. This study provides a GIS-based spatial characterization of 623 coal refuse sites in West Virginia, evaluating their potential for ecological and economic restoration. Stream loss was quantified through digitization of historic topographic maps, revealing over 194,500 meters of historically present streams lost, with 105,000 meters directly buried beneath refuse piles. Volume estimation using empirical Bayesian kriging and modern LiDAR surfaces indicated approximately 1.96 billion cubic meters of deposited refuse. Land cover analysis based on NAIP-derived classification showed that 48% of the total site area was forested, while 15% remained barren. These results highlight not only the environmental footprint of legacy refuse deposits but also their potential for restoration through stream daylighting and beneficial reuse. Emerging opportunities, including rare earth element recovery and mitigation banking, may offer economic incentives to facilitate reclamation. This work establishes a spatial framework to prioritize refuse sites for restoration, balancing ecological uplift with feasible material management strategies.

58 GEOSCIENCES↗
DOE OSTI2023

Bayesian Monte Carlo Evaluation of Imperfect (n, 233 U) Data and Model

Conventional nuclear data evaluation methods using generalized linear least squares make the following assumptions: prior and posterior probability distribution functions (PDFs) of all model parameters and data are normal (Gaussian); the linear approximation is sufficiently accurate to minimize the cost function (even for nonlinear models); the model (e.g., of neutron cross section) and experimental data (including covariance data) are without defect and prior PDFs of parameters and measured data are known perfectly. Neglect of covariance between model parameters and measured data in conventional evaluations contributes to imperfections. These assumptions are inherent to the generalized linear least squares minimization method commonly used for resolved resonance region neutron cross section evaluations but are often not justified due to the presence of non-normal PDFs, nonlinear models (e.g., R-matrix formalism), and inherent imperfections in data and models (e.g., imperfect covariance data). Here, these assumptions are removed in a mathematical framework of Bayes’ theorem, which is implemented using the Metropolis-Hastings Monte Carlo method. Most importantly, new parameters are introduced to parameterize discrepancies between the theoretical model and measured data to quantify judgement about discrepancies or imperfections in a reproducible manner. An evaluation of 233U in the eV region using the ENDF-B/VIII.0 library and transmission data (Guber et al.) is presented, and posterior parameters are compared to those obtained by conventional evaluation methods. This example illustrates the effects of removing the most harmful assumption: that of model-data perfection.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Unlocking hidden information in sparse small-angle neutron scattering measurements

Hypothesis Small-Angle Neutron Scattering (SANS) is a powerful technique for studying soft matter systems such as colloids, polymers, and lyotropic phases, providing nanoscale structural insights. However, its effectiveness is limited by low neutron flux, leading to long acquisition times and noisy data. Here, we hypothesize that Bayesian statistical inference using Gaussian Process Regression (GPR) can reconstruct high-fidelity scattering data from sparse measurements by leveraging intensity smoothness and continuity. Experiments and Simulations The method was benchmarked computationally and validated through SANS experiments on various soft matter systems, including wormlike micelles, colloidal suspensions, polymeric structures, and lyotropic phases. GPR-based inference was applied to both experimental and synthetic data to evaluate its effectiveness in noise reduction and intensity reconstruction. Findings GPR significantly enhances SANS data quality and therefore reducing measurement times by up to two orders of magnitude. This cost-effective approach maximizes experimental efficiency, enabling high-throughput studies and real-time monitoring of dynamic systems. It is particularly beneficial for weakly scattering and time-sensitive studies. Beyond SANS, this framework applies to other low-SNR techniques, including laboratory-based small-angle X-ray scattering and various dynamical scattering methods. Furthermore, it offers transformative potential for compact neutron sources, enhancing their viability for structural analysis in resource-limited settings.

Small angle neutron scattering↗
NASA NTRS1988

Probabilistic system identification in the time domain

The objective of system identification is to determine reliable dynamical models of a structure by systematically using its measured excitation and response. It brings together in an integrated fashion, experimental, analytical, and computational techniques in structural dynamics. Areas of application for system identification include the following: (1) Model Evaluation--assessing assumptions (linearity and equivalent viscous damping) and techniques (finite-element modeling) used to construct theoretical models of a structure; (2) Model Improvement--updating of a theoretical model to enable more accurate response predictions for possible future loads on the structure, or for control of the structure; (3) Empirical Modelling--developing empirical relationships (nonlinear models) or empirical parameter values (modal damping) because the present state of the art does not provide theoretical results; and (4) Damage Detection and Assessment--continual or episodic updating of a structural model through vibration monitoring to detect and locate any structural damage. It can be argued that since the construction or modification of models using test data is subject to inherent uncertainties, the above problems should be properly treated within a Bayesian probabilistic framework. Such a methodology is presented which allows the precision of the estimates of the model parameters to be computed. It also leads to a guiding principle in applications. Namely, when selecting a single model from a given class of models, one should take the most probable model in the class based on the experimental data. Practical applications of this principle are given which are based on the utilization of measured seismic motions in large civil structures. Examples include the application of a computer program MODE-ID to identify modal properties directly from seismic excitation and response time histories from a nine-story steel-frame building at JPL and from a freeway overpass bridge.

Beck, James L.↗
DOE OSTI2024

Structural mode coupling in perovskite oxides using hypothesis-driven active learning

Abstract Finding the ground-state structure with minimum energy is paramount to designing any material. In ABO 3 -type perovskite oxides with Pnma symmetry, the lowest energy phase is driven by an inherent trilinear coupling between the two primary order parameters such as rotation and tilt with antiferroelectric displacement of the A-site cations as established via hybrid improper ferroelectric mechanism. Conventionally, finding the relevant mode coupling driving phase transition requires performing first-principles calculations which is computationally time-consuming as well as expensive. It involves following an intuitive iterative hit and trial method of (a) adding two or multiple mode vectors, followed by (b) evaluating which combination would lead to the ground-state energy. In this study, we show how a hypothesis-driven active learning framework can identify suitable mode couplings within the Landau free energy expansion with minimal information on amplitudes of modes for a series of double perovskite oxides with A-site layered, columnar and rocksalt ordering. This scheme is expected to be applicable universally for understanding atomistic mechanisms derived from various structural mode couplings behind functionalities, for e.g. polarization, magnetization and metal–insulator transitions.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Parallelizing autotuning for HPC applications: Unveiling the potential of the speculation strategy in Bayesian optimization

In the exascale computing era, tuning High-Performance Computing (HPC) applications has become a significant computational challenge. Although Bayesian optimization (BO) has emerged as a promising tool for HPC performance tuning, the BO workflow is inherently sequential (i.e., one function evaluation at a time) and cannot leverage the huge amount of parallel resources present in modern supercomputers, resulting in a considerable underutilization of their computational capabilities. This paper explores the trade-off between search quality and parallelism in BO, investigating a diverse set of methods. Building upon both previous approaches from the literature and novel methodologies introduced in this work, our study provides a deep analysis to accelerate BO performance tuning. By examining a set of synthetic functions and practical HPC applications, our exploration analyzes the interaction among various BO methods for parallelization, the quantity of parallel resources, the runtime distribution of target HPC applications, and the costs associated with different search orchestration mechanisms that have been overlooked in previous studies. Compared to sequential BO, our novel methodology achieves comparable quality while demonstrating robust scalability in search time as the amount of parallel resources increases; it also outperforms a state-of-the-art tuner, which supports parallelization, achieving up to 3.67x faster search time. We provide high-value insights for practitioners seeking to leverage the power of parallel computing for efficient HPC application tuning. Additionally, to further assist researchers in accelerating the performance tuning of their HPC applications, we provide an extension of an existing open-source tuning framework that incorporates our methods.

Bayesian optimization↗
DOE OSTI2022

Bayesian Seismoacoustic Source Location: Acoustic Approach

Seismic waves and infrasound are key technologies in the International Monitoring System (IMS) to monitor explosive events in the solid Earth and atmosphere. Energetic man-made or natural events (e.g., chemical/nuclear explosions, volcanic eruptions, and earthquakes) near the Earth’s surface produce both ground motion and atmospheric pressure disturbances which propagate as seismic waves and infrasound, respectively. Seismic waves have been generally used to detect and identify underground and near-surface events (Myers, et. al., 2007), and infrasound are sensitive to events near the surface or in the atmosphere (Modrak et. al., 2010). Due to their different sensitivities to events, they can complement to each other to improve the event detection and discrimination. The framework of joint seismoacoustic event location has recently reviewed by theoretical research (Koch and Arrowsmith, 2019). Although the early applications showed promising results to improve the accuracy of event location, their application were still limited to a small set of events selected to prove the concepts, and practical capability of the method for operational purpose is not fully evaluated with data. Our final goal is to apply the method of seismoacoustic event location to a larger set of events and evaluate its applicability for operational event location in practice. To that end, we focus on developing and verifying acoustic source location method in this study.

58 GEOSCIENCES↗
DOE OSTI2024

Projective Integral Updates for High-Dimensional Variational Inference

Variational inference is an approximation framework for Bayesian inference that seeks to improve quantified uncertainty in predictions by optimizing a simplified distribution over parameters to stand in for the full posterior. Capturing model variations that remain consistent with training data enables more robust predictions by reducing parameter sensitivity. This work introduces a fixed-point optimization for variational inference that is applicable when every feasible log density can be expressed as a linear combination of functions from a given basis. In such cases, the optimizer becomes a fixed-point of projective integral updates. When the basis spans univariate quadratics in each parameter, the feasible distributions are Gaussian mean-fields and the projective integral updates yield quasi-Newton variational Bayes (QNVB). Other bases and updates are also possible. Since these updates require high-dimensional integration, this work begins by proposing an efficient quasirandom sequence of quadratures for mean-field distributions. Each iterate of the sequence contains two evaluation points that combine to correctly integrate all univariate quadratic functions and, if the mean-field factors are symmetric, all univariate cubics. More importantly, averaging results over short subsequences achieves periodic exactness on a much larger space of multivariate polynomials of quadratic total degree. The corresponding variational updates require four loss evaluations with standard (not second-order) backpropagation to eliminate error terms from over half of all multivariate quadratic basis functions. Furthermore, this integration technique is motivated by first proposing stochastic blocked mean-field quadratures, which may be useful in other contexts. A PyTorch implementation of QNVB allows for better control over model uncertainty during training than competing methods. Experiments demonstrate superior generalizability for multiple learning problems and architectures.

Gaussian mean-field↗
NASA NTRS2021

Liana optical traits increase tropical forest albedo and reduce ecosystem productivity

Lianas are a key growth form in tropical forests. Their lack of self-supporting tissues and their vertical position on top of the canopy make them strong competitors of resources. A few pioneer studies have shown that liana optical traits differ on average from those of colocated trees. Those trait discrepancies were hypothesized to be responsible for the competitive advantage of lianas over trees. Yet, in the absence of reliable modelling tools, it is impossible to unravel their impact on the forest energy balance, light competition, and on the liana success in Neotropical forests. To bridge this gap, we performed a meta-analysis of the literature to gather all published liana leaf optical spectra, as well as all canopy spectra measured over different levels of liana infestation. We then used a Bayesian data assimilation framework applied to two radiative transfer models (RTMs) covering the leaf and canopy scales to derive tropical tree and liana trait distributions, which finally informed a full dynamic vegetation model. According to the RTMs inversion, lianas grew thinner, more horizontal leaves with lower pigment concentrations. Those traits made the lianas very efficient at light interception and significantly modified the forest energy balance and its carbon cycle. While forest albedo increased by 14% in the shortwave, light availability was reduced in the understorey (−30% of the PAR radiation) and soil temperature decreased by 0.5°C. Those liana-specific traits were also responsible for a significant reduction of tree (−19%) and ecosystem (−7%) gross primary productivity (GPP) while lianas benefited from them (their GPP increased by +27%). This study provides a novel mechanistic explanation to the increase in liana abundance, new evidence of the impact of lianas on forest functioning, and paves the way for the evaluation of the large-scale impacts of lianas on forest biogeochemical cycles.

Félicien Meunier↗
DOE OSTI2022

Descriptor Aided Bayesian Optimization for Many-Level Qualitative Variables With Materials Design Applications

Abstract Engineering design often involves qualitative and quantitative design variables, which requires systematic methods for the exploration of these mixed-variable design spaces. Expensive simulation techniques, such as those required to evaluate optimization objectives in materials design applications, constitute the main portion of the cost of the design process and underline the need for efficient search strategies—Bayesian optimization (BO) being one of the most widely adopted. Although recent developments in mixed-variable Bayesian optimization have shown promise, the effects of dimensionality of qualitative variables have not been well studied. High-dimensional qualitative variables, i.e., with many levels, impose a large design cost as they typically require a larger dataset to quantify the effect of each level on the optimization objective. We address this challenge by leveraging domain knowledge about underlying physical descriptors, which embody the physics of the underlying physical phenomena, to infer the effect of unobserved levels that have not been sampled yet. We show that physical descriptors can be intuitively embedded into the latent variable Gaussian process approach—a mixed-variable GP modeling technique—and used to selectively explore levels of qualitative variables in the Bayesian optimization framework. This physics-informed approach is particularly useful when one or more qualitative variables are high dimensional (many-level) and the modeling dataset is small, containing observations for only a subset of levels. Through a combination of mathematical test functions and materials design applications, our method is shown to be robust to certain types of incomplete domain knowledge and significantly reduces the design cost for problems with high-dimensional qualitative variables.

Engineering↗