Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “optimal accuracy”

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 667 records · Page 37

Optimal Quantum Phase Estimation

By using a systematic optimization approach, we determine quantum states of light with definite photon number leading to the best possible precision in optical two-mode interferometry. Our treatment takes into account the experimentally relevant situation of photon losses. Our results thus reveal the benchmark for precision in optical interferometry. Although this boundary is generally worse than the Heisenberg limit, we show that the obtained precision beats the standard quantum limit, thus leading to a significant improvement compared to classical interferometers. We furthermore discuss alternative states and strategies to the optimized states which are easier to generate at the cost of only slightly lower precision.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

An optimal merging technique for high-resolution precipitation products

Precipitation products are currently available from various sources at higher spatial and temporal resolution than any time in the past. Each of the precipitation products has its strengths and weaknesses in availability, accuracy, resolution, retrieval techniques and quality control. By merging the precipitation data obtained from multiple sources, one can improve its information content by minimizing these issues. However, precipitation data merging poses challenges of scale-mismatch, and accurate error and bias assessment. In this paper we present Optimal Merging of Precipitation (OMP), a new method to merge precipitation data from multiple sources that are of different spatial and temporal resolutions and accuracies. This method is a combination of scale conversion and merging weight optimization, involving performance-tracing based on Bayesian statistics and trend-analysis, which yields merging weights for each precipitation data source. The weights are optimized at multiple scales to facilitate multiscale merging and better precipitation downscaling. Precipitation data used in the experiment include products from the 12-km resolution North American Land Data Assimilation (NLDAS) system, the 8-km resolution CMORPH and the 4-km resolution National Stage-IV QPE. The test cases demonstrate that the OMP method is capable of identifying a better data source and allocating a higher priority for them in the merging procedure, dynamically over the region and time period. This method is also effective in filtering out poor quality data introduced into the merging process.

54 ENVIRONMENTAL SCIENCES↗

MADE3D: Enabling the next generation of high-torque density wind generators by additive design and 3D printing

Direct-drive wind turbine generators are increasing in popularity, thanks to recent project developments—especially offshore, where reliability and efficiency are major cost drivers. Yet, high capital costs are forcing many original equipment manufacturers to consider lightweight, high-torque density generators for next-generation multi-megawatt turbines that may be difficult to realize by traditional design or manufacturing methods. In this study, we present a new design framework enabled by advanced machine learning and multimaterial additive manufacturing to perform a magnetic topology optimization that maximizes the torque per rotor active mass for a 15-megawatt direct-drive permanent magnet wind generator. A comparison of the proposed approach against conventional topology optimization demonstrated a significant increase in computational efficiency and accuracy in performance predictions. Results using single and multimaterial compositions for rotor core and magnets identify a wider choice of 3D printable designs for a given specification. A hybrid combination of sintered and dysprosium-free polymer-bonded magnets shows good potential for torque performance by saving material costs up to 8.75%. More than 30% improvement in rotor torque densities is identified which can marginally improve the overall generator torque density. With the rapid evolution of multipowder deposition technolgies, this study can greatly inspire a new paradigm for design-driven manufacturing with novel material compositions and lightweight, low-cost, high-strength multimaterial geometries that were previously unexplored for direct-drive generators.

17 WIND ENERGY↗

Enhancing segmentation fairness through curriculum learning and progressive loss: a centralized and federated perspective on radiograph analysis

Bias in medical image segmentation can lead to unequal performance across demographic subgroups, raising concerns about fairness and reliability in clinical AI systems. While deep learning models have achieved high segmentation accuracy, ensuring equitable performance across race and gender remains a significant challenge, particularly in privacy-sensitive healthcare environments. This study investigates fairness-aware medical image segmentation for hip and knee radiographs using deep learning models evaluated in both centralized and Federated Learning (FL) settings. We introduce Curriculum Learning (CL) strategies and Progressive Loss (PL) functions to regulate sample difficulty during training. In addition, we propose two novel fairness-oriented federated learning algorithms, Federated Intersection over Union (FedIoU) and Federated Intersection over Union with Outlier Analysis (FedIoUoutlier). Experiments are conducted using multiple segmentation backbones and simulated multi-site data partitions derived from the Osteoarthritis Initiative dataset. Model performance is evaluated using Intersection over Union (IoU), IoU standard deviation, Skewed Error Ratio (SER), and Min-Max Disparity across race and gender subgroups. Statistical significance was verified using paired t-tests to compare per-sample IoU performance against baseline configurations. Across both hip and knee segmentation tasks, curriculum learning and progressive loss strategies consistently improved segmentation accuracy and reduced demographic performance disparities in centralized training. In federated settings, fairness-aware aggregation further enhanced performance. Notably, FedIoUoutlier combined with balanced curriculum learning and tiered progressive loss achieved the highest mean IoU while yielding the lowest SER and Min-Max Disparity, indicating improved fairness without sacrificing accuracy. In several configurations, federated models matched or exceeded the performance of optimized centralized models, with statistically significant improvements in per-sample IoU over baseline configurations. The results demonstrate that structured training strategies and fairness-aware federated aggregation can jointly improve accuracy, stability, and demographic fairness in medical image segmentation. By integrating curriculum learning, progressive loss, and novel FL algorithms, this work provides a practical pathway toward equitable and privacy-preserving AI systems for medical imaging.

97 MATHEMATICS AND COMPUTING↗

Effect of Target Location on Dynamic Visual Acuity During Passive Horizontal Rotation

The vestibulo-ocular reflex (VOR) generates eye rotation to compensate for potential retinal slip in the specific plane of head movement. Dynamic visual acuity (DVA) has been utilized as a functional measure of the VOR. The purpose of this study was to examine changes in accuracy and reaction time when performing a DVA task with targets offset from the plane of rotation, e.g. offset vertically during horizontal rotation. Visual acuity was measured in 12 healthy subjects as they moved a hand-held joystick to indicate the orientation of a computer-generated Landolt C "as quickly and accurately as possible." Acuity thresholds were established with optotypes presented centrally on a wall-mounted LCD screen at 1.3 m distance, first without motion (static condition) and then while oscillating at 0.8 Hz (DVA, peak velocity 60 deg/s). The effect of target location was then measured during horizontal rotation with the optotypes randomly presented in one of nine different locations on the screen (offset up to 10 deg). The optotype size (logMar 0, 0.2 or 0.4, corresponding to Snellen range 20/20 to 20/50) and presentation duration (150, 300 and 450 ms) were counter-balanced across five trials, each utilizing horizontal rotation at 0.8 Hz. Dynamic acuity was reduced relative to static acuity in 7 of 12 subjects by one step size. During the random target trials, both accuracy and reaction time improved proportional to optotype size. Accuracy and reaction time also improved between 150 ms and 300 ms presentation durations. The main finding was that both accuracy and reaction time varied as a function of target location, with greater performance decrements when acquiring vertical targets. We conclude that dynamic visual acuity varies with target location, with acuity optimized for targets in the plane of motion. Both reaction time and accuracy are functionally relevant DVA parameters of VOR function.

Appelbaum, Meghan↗

Data-driven modeling of coarse mesh turbulence for reactor transient analysis using convolutional recurrent neural networks

Advanced nuclear reactors often exhibit complex thermal-fluid phenomena during transients. To accurately capture such phenomena, a coarse-mesh three-dimensional (3-D) modeling capability is desired for modern nuclear-system code. In the coarse-mesh 3-D modeling of advanced-reactor transients that involve flow and heat transfer, accurately predicting the turbulent viscosity is a challenging task that requires an accurate and computationally efficient model to capture the unresolved fine-scale turbulence. In this work, we propose a data-driven coarse-mesh turbulence model based on local flow features for the transient analysis of thermal mixing and stratification in a sodium-cooled fast reactor. The model has a coarse mesh setup to ensure computational efficiency, while it is trained by fine-mesh computational fluid dynamics (CFD) data to ensure accuracy. A novel neural network architecture, combining a densely connected convolutional network and a long-short-term-memory network, is developed that can efficiently learn from the spatial temporal CFD transient simulation results. The neural network model was trained and optimized on a loss-of flow transient and demonstrated high accuracy in predicting the turbulent viscosity field during the whole transient. The trained model's generalization capability was also investigated on two other transients with different inlet conditions. The study demonstrates the potential of applying the proposed data-driven approach to support the coarse-mesh multi-dimensional modeling of advanced reactors.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Respiratory motion sampling in 4DCT reconstruction for radiotherapy

Purpose: Phase-based and amplitude-based sorting techniques are commonly used in four-dimensional CT (4DCT) reconstruction. However, effect of these sorting techniques on 4D dose calculation has not been explored. In this study, the authors investigated a candidate 4DCT sorting technique by comparing its 4D dose calculation accuracy with that for phase-based and amplitude-based sorting techniques.Method: An optimization model was formed using organ motion probability density function (PDF) in the 4D dose convolution. The objective function for optimization was defined as the maximum difference between the expected 4D dose in organ of interest and the 4D dose calculated using a 4DCT sorted by a candidate sampling method. Sorting samples, as optimization variables, were selected on the respiratory motion PDF assessed during the CT scanning. Breathing curves obtained from patients' 4DCT scanning, as well as 3D dose distribution from treatment planning, were used in the study. Given the objective function, a residual error analysis was performed, and k-means clustering was found to be an effective sampling scheme to improve the 4D dose calculation accuracy and independent with the patient-specific dose distribution. Results: Patient data analysis demonstrated that the k-means sampling was superior to the conventional phase-based and amplitude-based sorting and comparable to the optimal sampling results. For phase-based sorting, the residual error in 4D dose calculations may not be further reduced to an acceptable accuracy after a certain number of phases, while for amplitude-based sorting, k-means sampling, and the optimal sampling, the residual error in 4D dose calculations decreased rapidly as the number of 4DCT phases increased to 6.Conclusion: An innovative phase sorting method (k-means method) is presented in this study. The method is dependent only on tumor motion PDF. It could provide a way to refine the phase sorting in 4DCT reconstruction and is effective for 4D dose accumulation. Optimized sorting techniques could achieve acceptable residuals (less than 0.5% of the prescription dose) using 6 sorting samples, which is much better than amplitude-based or phase-based sorting. Further increase in sorting phase number exceeding 6 or more may not be necessary when using the k-means sampling or optimal sampling points.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Optimal control model of technology transition.

This paper discusses the use of optimization software to solve an optimal control problem arising in the modeling of technology transition. We set up a series of increasingly complex models with such features as learning-by-doing, adjustment cost, and capital investment. The models are written in continuous time and then discretized by using different methods to transform them into large-scale nonlinear programs. We use a modeling language and numerical optimization methods to solve the optimization problem. Our results are consistent with findings in the literature and highlight the impact the discretization choice has on the solution and accuracy. The primary goal of this paper is to develop solution methods for use in policy work. To do this, we will need to expand the model to include realistic features such as additional technologies and capital conversion. Our modeling approach allows us to add new technologies with minimal effort. For example, Figure 5 presents the solution to the capital model that includes a third technology, say, nuclear fusion, that is emission free but even more expensive (A{sub 3} > A{sub n}). We observe that industry first transitions to the intermediate technology but eventually adopts the cleanest one. We note that our solutions have few areas of large curvature and are relatively smooth elsewhere. On such smooth segments, we can achieve adequate precision with relatively large periods, which means fewer variables and increased computation speed. But we still need a small step length during the transition period. Adaptive mesh refinement methods have been developed to automatically identify the areas where extra precision is needed. Betts (2001) describes how the computation time can be reduced by a factor of 2.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

On the Rapid Calculation of Binding Affinities for Antigen and Antibody Design and Affinity Maturation Simulations

The accurate and efficient calculation of protein-protein binding affinities is an essential component in antibody and antigen design and optimization, and in computer modeling of antibody affinity maturation. Such calculations remain challenging despite advances in computer hardware and algorithms, primarily because proteins are flexible molecules, and thus, require explicit or implicit incorporation of multiple conformational states into the computational procedure. The astronomical size of the amino acid sequence space further compounds the challenge by requiring predictions to be computed within a short time so that many sequence variants can be tested. In this study, we compare three classes of methods for antibody/antigen (Ab/Ag) binding affinity calculations: (i) a method that relies on the physical separation of the Ab/Ag complex in equilibrium molecular dynamics (MD) simulations, (ii) a collection of 18 scoring functions that act on an ensemble of structures created using homology modeling software, and (iii) methods based on the molecular mechanics-generalized Born surface area (MM-GBSA) energy decomposition, in which the individual contributions of the energy terms are scaled to optimize agreement with the experiment. When applied to a set of 49 antibody mutations in two Ab/HIV gp120 complexes, all of the methods are found to have modest accuracy, with the highest Pearson correlations reaching about 0.6. In particular, the most computationally intensive method, i.e., MD simulation, did not outperform several scoring functions. The optimized energy decomposition methods provided marginally higher accuracy, but at the expense of requiring experimental data for parametrization. Within each method class, we examined the effect of the number of independent computational replicates, i.e., modeled structures or reinitialized MD simulations, on the prediction accuracy. We suggest using about ten modeled structures for scoring methods, and about five simulation replicates for MD simulations as a rule of thumb for obtaining reasonable convergence. We anticipate that our study will be a useful resource for practitioners working to incorporate binding affinity calculations within their protein design and optimization process.

59 BASIC BIOLOGICAL SCIENCES↗

Surrogate modeling and optimization of the leaching process in a rare earth elements recovery plant

Critical minerals (CMs) and Rare Earth Elements (REEs) play a vital role in crucial infrastructure technologies such as renewable energy generation and batteries. Recovering them from waste materials has recently been found to significantly reduce environmental impact and supply chain costs related to these materials. In this work, we investigate surrogate modeling techniques aimed to simplify the modeling, simulation, and optimization of the leaching processes involved in CM and REE recovery flowsheets. As there is currently a lack of systematic studies on this topic, we perform extensive computational testing to ascertain which surrogate models are easier to construct and offer high predictive accuracy. Further, our results suggest that sparse quadratic models balance predictive accuracy and computational efficiency. Training and using these surrogates for global optimization of the leaching process requires two orders of magnitude fewer measurements and is up to four orders of magnitude faster than optimizing the original simulation using equation-oriented optimization or derivative-free optimization.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

SU-E-T-255: Optimized Supine Craniospinal Irradiation with Image-Guided and Field Matched Beams

Purpose: Conventional craniospinal irradiation (CSI) challenges include dose inhomogeneity at field junctions and position uncertainty due to the field divergence, particular for the two spinal fields. Here we outline a new supine CSI technique to address these difficulties. Methods: Patient was simulated in supine position. The cranial fields had isocenter at C2/C3 vertebral and were matched with 1st spinal field. Their inferior border was chosen to avoid the shoulder, as well as chin from the 1st spine field. Their collimator angles were dependent on asymmetry jaw setting of the 1st spinal field. With couch rotation, the spinal field gantry angles were adjusted to ensure, the inferior border of 1st and superior border of 2nd spinal fields were perpendicular to the table top. The radio-opaque wire position for the spinal junction was located initially by the light field from an anterior setup beam, and was finalized by the portal imaging of the 1st spinal field. With reference to the spinal junction wire, the fields were matched by positioning the isocenter of the 2nd spinal field. A formula was derived to optimize supine CSI treatment planning, by utilizing the relationship among the Yjaw setting, the spinal field gantry angles, cranial field collimator angles, and the spinal field isocenters location. The plan was delivered with portal imaging alignment for the both cranial and spinal junctions. Results: Utilizing this technique with matching beams, and conventional technique such as feathering and forwarding planning, a homogenous dose distribution was achieved throughout the entire CSI treatment volume including the spinal junction. Placing the spinal junction wire visualized in both spinal portals, allows for precise determination and verification of the appropriate match line of the spine fields. Conclusion: This technique of optimization supine CSI achieved a homogenous dose distributions and patient localization accuracy with image-guided and matched beams.

60 APPLIED LIFE SCIENCES↗

Reliability-based design optimization using efficient global reliability analysis.

Finding the optimal (lightest, least expensive, etc.) design for an engineered component that meets or exceeds a specified level of reliability is a problem of obvious interest across a wide spectrum of engineering fields. Various methods for this reliability-based design optimization problem have been proposed. Unfortunately, this problem is rarely solved in practice because, regardless of the method used, solving the problem is too expensive or the final solution is too inaccurate to ensure that the reliability constraint is actually satisfied. This is especially true for engineering applications involving expensive, implicit, and possibly nonlinear performance functions (such as large finite element models). The Efficient Global Reliability Analysis method was recently introduced to improve both the accuracy and efficiency of reliability analysis for this type of performance function. This paper explores how this new reliability analysis method can be used in a design optimization context to create a method of sufficient accuracy and efficiency to enable the use of reliability-based design optimization as a practical design tool.

97 MATHEMATICS AND COMPUTING↗

Frequency-Domain Deconvolution for Flight Dynamics Applications

A deconvolution method is presented for estimating input data from measured output data and a model of the dynamic process involved. The method uses an optimal Wiener filter for separating the measured data into signal and noise components, and a high-accuracy Fourier transform for inverting the model dynamics in the frequency domain. The method is an extension of optimal Fourier smoothing, and uses a technique to enhance the contrast between the signal and noise spectra in designing the Wiener filter. The deconvolution method was applied to simulation and flight test data for the purposes of removing unwanted distortions introduced by signal-conditioning filters and sensor dynamics, and for reconstructing turbulence inputs from measured sensor data. Results indicated hat the method performs well given good signal-to-noise levels and accurate models of the dynamic process.

Grauer, Jared A.↗

Sensor Networks for Detecting Toxic Releases in Buildings

Sudden releases of a toxic agent indoors can cause immediate and long-term harm to occupants. In order to protect building occupants from such threats, it is necessary to have a robust air monitoring system that can detect, locate, and characterize accidental or deliberate toxic gas releases. However, developing such a system is complicated by several requirements, in particular the need to operate in real-time. This task is further complicated when monitoring sensors are prone to false positive and false negative readings. We report on work towards developing an indoor monitoring system that is robust even in the presence of poor quality sensor data. The algorithm, named BASSET, combines deterministic modeling and Bayesian statistics to join prior knowledge of the contaminant transport in the building with real-time sensor information. We evaluate BASSET across several data sets, which vary in sensor characteristics such as accuracy, response time, and trigger level. Our results suggest that optimal designs are not always intuitive. For example, a network comprised of slower but more accurate sensors may locate the contaminant source more quickly than a network with faster but less accurate sensors.

54 ENVIRONMENTAL SCIENCES↗

Volumetrically Absorbing Thermal Insulator (VATI) for High-temperature Receivers

This project seeks to exploit volumetric absorption of concentrated solar irradiation in a thin, high-temperature, thermally conductive medium. High efficiency thermal conversion is achieved by volumetric absorption of both incoming solar irradiation (short wavelength) and emitted irradiation (long wavelength), demonstrating a volumetrically absorbing thermally insulating (VATI) effect. The thermalized energy is conducted to the back wall, enclosing the working fluid (molten salt or supercritical CO 2 ). Reliance on conductive thermal transport requires a thermally conductive medium, in the absence of which large temperature gradients drive losses due to emission. The project’s outcomes are important to realize a cost-effective approach to reduce optical and thermal losses from CSP receivers at high temperatures (720°C). High-temperature stable and commercially available porous SiC structures (open-cell foams) were explored as a volumetrically absorbing and thermally insulating layer. To the best of our knowledge, the current state-of-the-art for industrially deployed coatings is Pyromark. However, Pyromark suffers degradation at 700+°C and diurnal temperature changes. This necessitates periodic recoating, and the downtime increases the overall levelized cost of energy (LCOE) production. On the other hand, contemporary research activities have generated significant advances in the development of selective emitter coatings, which present challenges with costs, scalability, and stability. The pursued approach alleviates these concerns by developing a receiver that utilizes the inherent structure of high-temperature stable porous materials to enable robust and cost-effective receivers which require no periodic maintenance downtimes. The overall goal of the project is to experimentally demonstrate a figure of merit (FOM) of 0.92 at a temperature of 720°C and solar irradiation of 1000x concentration with porous receivers. The relevant crystallographic (phase) optical and thermal properties of porous SiC were first characterized. Second, the 3-D geometry of the porous SiC was analyzed using micro-X-ray tomography and converted to CAD data using image processing analysis. Utilizing this 3-D geometry and relevant optical/thermal properties, Monte Carlo- Ray Tracing (MCRT) analysis was performed to extract important parameters governing solar-thermal energy conversion such as extinction coefficient (β, 1/m), scattering albedo (ω) and the scattering phase function (Φ). These properties were then integrated into an in-house radiative and conduction transport model to solve for temperature and transport fluxes characterizing the solar-thermal energy conversion. This model was utilized to predict the FOM for various SiC porous geometries and to identify the highest possible FOM. Testing of the optimized porous structures will be accomplished with a custom-built high-accuracy (< ±4%) FOM measurement test-stand and a 1000x solar concentrator. When neglecting convective losses and resistance at the open boundary and the back wall, the optimized SiC foam leads to a FOM of 0.84. This FOM does not surpass the performance of Pyromark 2500. Yet, conversely to Pyromark 2500, SiC is stable at high temperatures and does not degrade over time. Also, it is possible to boost the FOM of SiC by engineering its effective thermal conductivity and its scattering albedo. A FOM of 0.92 is predicted for an optimized foam by accounting for convective losses and resistance at the open boundary and the back wall. This largely exceeds the FOM of Pyromark 2500 (~0.86) predicted by neglecting convective losses and resistance. Therefore, further engineering of the foam could lead to unprecedented FOMs.

14 SOLAR ENERGY↗

Fourier-DeepONet: Fourier-enhanced deep operator networks for full waveform inversion with improved accuracy, generalizability, and robustness

In this article, full waveform inversion (FWI) infers the subsurface structure information from seismic waveform data by solving a non-convex optimization problem. Data-driven FWI has been increasingly studied with various neural network architectures to improve accuracy and computational efficiency. Nevertheless, the applicability of pre-trained neural networks is severely restricted by potential discrepancies between the source function used in the field survey and the one utilized during training. Here, we develop a Fourier-enhanced deep operator network (Fourier-DeepONet) for FWI with the generalization of seismic sources, including the frequencies and locations of sources. Specifically, we employ the Fourier neural operator as the decoder of DeepONet, and we utilize source parameters as one input of Fourier-DeepONet, facilitating the resolution of FWI with variable sources. To test Fourier-DeepONet, we develop three new and realistic FWI benchmark datasets (FWI-F, FWI-L, and FWI-FL) with varying source frequencies, locations, or both. Our experiments demonstrate that compared with existing data-driven FWI methods, Fourier-DeepONet obtains more accurate predictions of subsurface structures in a wide range of source parameters. Moreover, the proposed Fourier-DeepONet exhibits superior robustness when handling data with Gaussian noise or missing traces and sources with Gaussian noise, paving the way for more reliable and accurate subsurface imaging across diverse real conditions.

42 ENGINEERING↗

High-Fidelity Adiabatic Passage by Composite Sequences of Chirped Pulses

We present a method for optimization of the technique of adiabatic passage between two quantum states by composite sequences of frequency-chirped pulses with specific relative phases: composite adiabatic passage (CAP). By choosing the composite phases appropriately the nonadiabatic losses can be canceled to any desired order with sufficiently long sequences, regardless of the nonadiabatic coupling. The values of the composite phases are universal for they do not depend on the pulse shapes and the chirp. The accuracy of the CAP technique and its robustness against parameter variations make CAP suitable for high-fidelity quantum information processing.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Enhancing Gaussian Process Surrogates for Optimization and Posterior Approximation via Random Exploration

This paper proposes novel noise-free Bayesian optimization strategies that rely on a random exploration step to enhance the accuracy of Gaussian process surrogate models. The new algorithms retain the ease of implementation of the classical GP-UCB algorithm, but the additional random exploration step accelerates their convergence, nearly achieving the optimal convergence rate. Furthermore, to facilitate Bayesian inference with intractable likelihoods, we propose to utilize optimization iterates for maximum a posteriori estimation to build a Gaussian process surrogate model for the unnormalized log-posterior density. We provide bounds for the Hellinger distance between the true and the approximate posterior distributions in terms of the number of design points. We demonstrate the effectiveness of our Bayesian optimization algorithms in nonconvex benchmark objective functions, in a machine learning hyperparameter tuning problem, and in a black-box engineering design problem. The effectiveness of our posterior approximation approach is demonstrated in two Bayesian inference problems for parameters of dynamical systems.

Bayesian inference↗