Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “probabilistic inference”

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 289 records · Page 16

A controlled-NOT gate for frequency-bin qubits

The realization of strong photon–photon interactions has presented an enduring challenge across photonics, particularly in quantum computing, where two-photon gates form essential components for scalable quantum information processing (QIP). While linear-optic schemes have enabled probabilistic entangling gates in spatio-polarization encoding, solutions for many other useful degrees of freedom remain missing. In particular, no two-photon gate for the important platform of frequency encoding has been experimentally demonstrated, due in large part to the additional challenges imparted by the mismatched wavelengths of the interacting photons. In this article, we design and implement an entangling gate for frequency-bin qubits, a coincidence-basis controlled-NOT (CNOT), using line-by-line pulse shaping and electro-optic modulation. With this being said, we extract a quantum unitary fidelity of 0.91 ± 0.01 via a parameter inference approach based on Bayesian machine learning, which enables accurate gate reconstruction from measurements in the two-photon computational basis alone. Thus, our CNOT imparts a single-photon frequency shift controlled by the frequency of another photon—an important capability in itself—and should enable new directions in fiber-compatible QIP.

42 ENGINEERING↗

Probabilistic picture for particle number densities in stretched tips of the branching Brownian motion

Abstract In the framework of a stochastic picture for the one-dimensional branching Brownian motion, we compute the probability density of the number of particles near the rightmost one at a time T , that we take very large, when this extreme particle is conditioned to arrive at a predefined position x T chosen far ahead of its expected position m T . We recover the previously conjectured fact that the typical number density of particles at a distance Δ to the left of the lead particle, when both Δ and are large, is smaller than the mean number density by a factor proportional to , where ζ is a constant that was so far undetermined. Our picture leads to an expression for the probability density of the particle number, from which a value for ζ may be inferred.

Physics↗

Methods for Determining Subsets of High Impact, Probabilistically Dependent Medical Conditions Represented in a Directed Graph

One of the longest standing questions in network theory is how a component influences other parts in the system, and how that role is affected when restricting the navigation through the network. The Katz score, one of many centrality measures created for this purpose, takes into account all possible walks through the network, penalizing each additional step in a walk by a scalar called the Katz parameter. This centrality measure often covers an infinite number of walks with infinite length. In this paper we identify the maximum path length which has influence on the Katz score. We ultimately provide guidance when deciding which Katz parameter to use as it depends on the path length of interest. We show how changing the Katz parameter affects the ranking of the vertices in some synthetic graphs as well as NASA's expert informed network of medical dependencies called the Susceptibility Inference Network (SIN).

Hunter Rehm↗

Emulation With Uncertainty Quantification of Regional Sea‐Level Change Caused by the Antarctic Ice Sheet

Abstract Projecting regional sea‐level change under various climate‐change scenarios typically involves running forward simulations of the Earth's gravitational, rotational and deformational (GRD) response to ice‐mass change, which requires substantial computational cost if applied to probabilistic frameworks requiring thousands to millions of samples. Here we build emulators of regional sea‐level change at 27 coastal locations, due to the GRD effects associated with future Antarctic Ice Sheet mass change over the 21st century. The emulators are evaluated against a numerical sea‐level model applied to an ensemble of ice‐sheet model simulations of the Antarctic Ice Sheet through 2100. We build a physics‐based emulator using a recent sensitivity kernel approach and compare it to machine learning based emulators (neural network and conditional variational autoencoder methods). In order to quantify uncertainty, we derive well‐calibrated prediction intervals for regional sea‐level change via split‐conformal inference and linear regression, and show that Monte Carlo dropout does not yield well‐calibrated uncertainties in this instance. We also demonstrate substantial gains in computational efficiency using both the physics‐based emulator and neural networks in comparison to the numerical model for the complete regional sea‐level solution. Overall, we find the physics‐based emulator modestly outperforms the machine learning emulators for this problem.

58 GEOSCIENCES↗

Reinforcement Learning via Gaussian Processes with Neural Network Dual Kernels

While deep neural networks (DNNs) and Gaussian Processes (GPs) are both popularly utilized to solve problems in reinforcement learning, both approaches feature undesirable drawbacks for challenging problems. DNNs learn complex non-linear embeddings, but do not naturally quantify uncertainty and are often data-inefficient to train. GPs infer posterior distributions over functions, but popular kernels exhibit limited expressivity on complex and high-dimensional data. Fortunately, recently discovered conjugate and neural tangent kernel functions encode the behavior of overparameterized neural networks in the kernel domain. We demonstrate that these kernels can be efficiently applied to regression and reinforcement learning problems by analyzing a baseline case study.We apply GPs with neural network dual kernels to solve reinforcement learning tasks for the first time. We demonstrate, using the well understood mountain-car problem, that GPs empowered with dual kernels perform at least as well as those using the conventional radial basis function kernel. Finally, we conjecture that by inheriting the probabilistic rigor of GPs and the powerful embedding properties of DNNs, GPs using NN dual kernels will empower future reinforcement learning models on difficult domains.

97 MATHEMATICS AND COMPUTING↗

A machine learning Automated Recommendation Tool for synthetic biology

Abstract Synthetic biology allows us to bioengineer cells to synthesize novel valuable molecules such as renewable biofuels or anticancer drugs. However, traditional synthetic biology approaches involve ad-hoc engineering practices, which lead to long development times. Here, we present the Automated Recommendation Tool (ART), a tool that leverages machine learning and probabilistic modeling techniques to guide synthetic biology in a systematic fashion, without the need for a full mechanistic understanding of the biological system. Using sampling-based optimization, ART provides a set of recommended strains to be built in the next engineering cycle, alongside probabilistic predictions of their production levels. We demonstrate the capabilities of ART on simulated data sets, as well as experimental data from real metabolic engineering projects producing renewable biofuels, hoppy flavored beer without hops, fatty acids, and tryptophan. Finally, we discuss the limitations of this approach, and the practical consequences of the underlying assumptions failing.

59 BASIC BIOLOGICAL SCIENCES↗

Combining biomarker and virus phylogenetic models improves HIV-1 epidemiological source identification

To identify and stop active HIV transmission chains new epidemiological techniques are needed. Here, we describe the development of a multi-biomarker augmentation to phylogenetic inference of the underlying transmission history in a local population. HIV biomarkers are measurable biological quantities that have some relationship to the amount of time someone has been infected with HIV. To train our model, we used five biomarkers based on real data from serological assays, HIV sequence data, and target cell counts in longitudinally followed, untreated patients with known infection times. The biomarkers were modeled with a mixed effects framework to allow for patient specific variation and general trends, and fit to patient data using Markov Chain Monte Carlo (MCMC) methods. Subsequently, the density of the unobserved infection time conditional on observed biomarkers were obtained by integrating out the random effects from the model fit. This probabilistic information about infection times was incorporated into the likelihood function for the transmission history and phylogenetic tree reconstruction, informed by the HIV sequence data. To critically test our methodology, we developed a coalescent-based simulation framework that generates phylogenies and biomarkers given a specific or general transmission history. Testing on many epidemiological scenarios showed that biomarker augmented phylogenetics can reach 90% accuracy under idealized situations. Under realistic within-host HIV-1 evolution, involving substantial within-host diversification and frequent transmission of multiple lineages, the average accuracy was at about 50% in transmission clusters involving 5–50 hosts. Realistic biomarker data added on average 16 percentage points over using the phylogeny alone. Using more biomarkers improved the performance. Shorter temporal spacing between transmission events and increased transmission heterogeneity reduced reconstruction accuracy, but larger clusters were not harder to get right. More sequence data per infected host also improved accuracy. We show that the method is robust to incomplete sampling and that adding biomarkers improves reconstructions of real HIV-1 transmission histories. The technology presented here could allow for better prevention programs by providing data for locally informed and tailored strategies.

60 APPLIED LIFE SCIENCES↗

VpROM: a novel variational autoencoder-boosted reduced order model for the treatment of parametric dependencies in nonlinear systems

Reduced Order Models (ROMs) are of considerable importance in many areas of engineering in which computational time presents difficulties. Established approaches employ projection-based reduction, such as Proper Orthogonal Decomposition. The limitation of the linear nature of such operators is typically tackled via a library of local reduction subspaces, which requires the assembly of numerous local ROMs to address parametric dependencies. Our work attempts to define a more generalisable mapping between parametric inputs and reduced bases for the purpose of generative modeling. We propose the use of Variational Autoencoders (VAEs) in place of the typically utilised clustering or interpolation operations, for inferring the fundamental vectors, termed as modes, which approximate the manifold of the model response for any and each parametric input state. The derived ROM still relies on projection bases, built on the basis of full-order model simulations, thus retaining the imprinted physical connotation. However, it additionally exploits a matrix of coefficients that relates each local sample response and dynamics to the global phenomena across the parametric input domain. The VAE scheme is utilised for approximating these coefficients for any input state. This coupling leads to a high-precision low-order representation, which is particularly suited for problems where model dependencies or excitation traits cause the dynamic behavior to span multiple response regimes. Moreover, the probabilistic treatment of the VAE representation allows for uncertainty quantification on the reduction bases, which may then be propagated to the ROM response. The performance of the proposed approach is validated on an open-source simulation benchmark featuring hysteresis and multi-parametric dependencies, and on a large-scale wind turbine tower characterised by nonlinear material behavior and model uncertainty.

Conditional VAEs↗

Variation-Resilient FeFET-Based In-Memory Computing Leveraging Probabilistic Deep Learning

Reliability issues stemming from device level nonidealities of nonvolatile emerging technologies like ferroelectric field-effect transistors (FeFETs), especially at scaled dimensions, cause substantial degradation in the accuracy of in-memory crossbar-based AI systems. Here, in this work, we present a variation-aware design technique to characterize the device level variations and to mitigate their impact on hardware accuracy employing a Bayesian neural network (BNN) approach. An effective conductance variation model is derived from the experimental measurements of cycle-to-cycle (C2C) and device-to-device (D2D) variations performed on FeFET devices fabricated using 28 nm high-k metal gate technology. The variations were found to be a function of different conductance states within the given programming range, which sharply contrasts earlier efforts where a fixed variation dispersion was considered for all conductance values. Such variation characteristics formulated for three different device sizes at different read voltages were provided as prior variation information to the BNN to yield a more exact and reliable inference. Near-ideal accuracy for shallow networks (MLP5 and LeNet models) on the MNIST dataset and limited accuracy decline by ~3.8%–16.1% for deeper AlexNet models on CIFAR10 dataset under a wide range of variations corresponding to different device sizes and read voltages, demonstrates the efficacy of our proposed device-algorithm co-design technique.

97 MATHEMATICS AND COMPUTING↗

Particle Filter Based Inference Testing

The primary intent of PAR-FIT (Particle Filter based Inference Testing) is to provide hard inductive evidence that a machine learning model is capable and proven for an individual test input. By examining training data used to form the underlying model functional correlation, an estimate of the reliability that a model will make the correct prediction can be made. The Sequential Probability Ratio Test is used to derive a qualitative evaluation for reliability based on hypothesis testing. The PAR-FIT framework achieves this by implementing a particle filter and the sequential probability ratio test algorithms on the machine learning model training data to determine relevancy of new individual test samples to the training dataset. The kernel function evaluates the local proximity and density of training data used to derive a prediction outcome. Particles are used to probabilistically determine which training data to evaluate for proximity. For test samples that are within a close proximity to and surrounded by multiple training data points, the evaluated reliability of the prediction is high. For test samples that are anomalies not represented by the training dataset, in low density data clusters, or are far from existing data points, the evaluated reliability is low as insufficient training evidence exists to suggest the model is capable of making the correct prediction. Sequential Probability Ratio Test is further used to determine when a hypothesis on whether a signal can be rejected or accepted for use. The ratio test collects sequence information from the particle filter to test whether the signal is anomalous or normal via hypothesis testing of the underlying distributions.

Chen, Edward [Idaho National Laboratory (INL), Ida↗

Characterizing Regional Methane Emissions from Natural Gas Liquids Unloading

A “bottom-up” probabilistic model was developed using engineering first-principles to quantify annualized throughput normalized methane emissions (TNME) from natural gas liquid unloading activities for 18 basins in the United States in 2016. For each basin, six discrete liquid-unloading scenarios are considered, consisting of combinations of well types (conventional and unconventional) and liquid-unloading systems (nonplunger, manual plunger lift, and automatic plunger lift). While previous work has advocated that operational malfunctions and abnormal process conditions explain the existence of super-emitters in the natural gas supply chain, this work finds that super-emitters can arise naturally due to variability in underlying component processes. Additionally, average cumulative methane emissions from liquids unloading, attributed to the natural gas supply chain, across all basins are ~4.8 times higher than those inferred from the 2016 Greenhouse Gas Reporting Program (GHGRP).

Zaimes, George G.↗

On the Probability of Occurrence of Extreme Space Weather Events

By virtue of their rarity, extreme space weather events, such as the Carrington event of 1859, are difficult to study, their rates of occurrence are difficult to estimate, and prediction of a specific future event is virtually impossible. Additionally, events may be extreme relative to one parameter but normal relative to others. In this study, we analyze several measures of the severity of space weather events (flare intensity, coronal mass ejection speeds, Dst, and greater than 30 MeV proton fluences as inferred from nitrate records) to estimate the probability of occurrence of extreme events. By showing that the frequency of occurrence scales as an inverse power of the severity of the event, and assuming that this relationship holds at higher magnitudes, we are able to estimate the probability that an event larger than some criteria will occur within a certain interval of time in the future. For example, the probability of another Carrington event (based on Dst less than - 850 nT) occurring within the next decade is approximately 12%. We also identify and address several limitations with this approach. In particular, we assume time stationarity, and thus, the effects of long-term space climate change are not considered. While this technique cannot be used to predict specific events, it may ultimately be useful for probabilistic forecasting.

Riley, Pete↗

EFIT-Prime: Probabilistic and physics-constrained reduced-order neural network model for equilibrium reconstruction in DIII-D

We introduce EFIT-Prime, a novel machine learning surrogate model for EFIT (Equilibrium FIT) that integrates probabilistic and physics-informed methodologies to overcome typical limitations associated with deterministic and ad hoc neural network architectures. EFIT-Prime utilizes a neural architecture search-based deep ensemble for robust uncertainty quantification, providing scalable and efficient neural architectures that comprehensively quantify both data and model uncertainties. Physically informed by the Grad–Shafranov equation, EFIT-Prime applies a constraint on the current density J tor and a smoothness constraint on the first derivative of the poloidal flux, ensuring physically plausible solutions. Furthermore, the spatial location of the diagnostics is explicitly incorporated in the inputs to account for their spatial correlation. Extensive evaluations demonstrate EFIT-Prime's accuracy and robustness across diverse scenarios, most notably showing good generalization on negative-triangularity discharges that were excluded from training. Timing studies indicate an ensemble inference time of 15 ms for predicting a new equilibrium, offering the possibility of plasma control in real-time, if the model is optimized for speed.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Design of Experiments, Model Calibration and Data Assimilation

This presentation provides an overview of emulation, calibration and experiment design for computer experiments. Emulation refers to building a statistical surrogate from a carefully selected and limited set of model runs to predict unsampled outputs. The standard kriging approach to emulation of complex computer models is presented. Calibration refers to the process of probabilistically constraining uncertain physics/engineering model inputs to be consistent with observed experimental data. An initial probability distribution for these parameters is updated using the experimental information. Markov chain Monte Carlo (MCMC) algorithms are often used to sample the calibrated parameter distribution. Several MCMC algorithms commonly employed in practice are presented, along with a popular diagnostic for evaluating chain behavior. Space-filling approaches to experiment design for selecting model runs to build effective emulators are discussed, including Latin Hypercube Design and extensions based on orthogonal array skeleton designs and imposed symmetry requirements. Optimization criteria that further enforce space-filling, possibly in projections of the input space, are mentioned. Designs to screen for important input variations are summarized and used for variable selection in a nuclear fuels performance application. This is followed by illustration of sequential experiment design strategies for optimization, global prediction, and rare event inference.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Bridging the Gap between Human Judgment and Automated Reasoning in Predictive Analytics

Events occur daily that impact the health, security and sustainable growth of our society. If we are to address the challenges that emerge from these events, anticipatory reasoning has to become an everyday activity. Strong advances have been made in using integrated modeling for analysis and decision making. However, a wider impact of predictive analytics is currently hindered by the lack of systematic methods for integrating predictive inferences from computer models with human judgment. In this paper, we present a predictive analytics approach that supports anticipatory analysis and decision-making through a concerted reasoning effort that interleaves human judgment and automated inferences. We describe a systematic methodology for integrating modeling algorithms within a serious gaming environment in which role-playing by human agents provides updates to model nodes and the ensuing model outcomes in turn influence the behavior of the human players. The approach ensures a strong functional partnership between human players and computer models while maintaining a high degree of independence and greatly facilitating the connection between model and game structures.

99 GENERAL AND MISCELLANEOUS↗

Machine Learning‐Assisted Microearthquake Location Workflow for Monitoring the Newberry Enhanced Geothermal System

Abstract Enhanced geothermal systems (EGS) offer a sustainable energy source but face challenges in accurately locating microearthquakes induced during reservoir stimulation. Locating these microearthquakes provides reliable feedback on the stimulation progress. Current deep learning methods for locating earthquakes require extensive data sets for training, which is problematic as detected microearthquakes are often limited. To address the scarcity of training data, we propose a practical workflow using probabilistic multilayer perceptron (PMLP) which predicts microearthquake locations from cross‐correlation time lags in waveforms. Utilizing a 3D velocity model of Newberry site derived from ambient noise interferometry, we generate numerous synthetic microearthquakes and 3D acoustic waveforms for PMLP training. Accurate synthetic tests prompt us to apply the trained network to the 2012 and 2014 stimulation field waveforms. To enhance the accuracy of source localization, we carefully handpick the P‐arrival times. Predictions on the 2012 stimulation data set show major microseismic activity at depths of 0.5–1.2 km, correlating with a known casing leakage scenario. In the 2014 data set, the majority of predictions concentrate at 2.0–2.9 km depths, consistent with results obtained from conventional physics‐based inversion, and align with the presence of natural fractures from 2.0 to 2.7 km. We validate our findings by comparing the synthetic and field picks, demonstrating a satisfactory match for the first arrivals. By combining the benefits of quick inference speeds and accurate location predictions, we demonstrate the feasibility of using realistic synthetic data set to locate microseismicity for EGS monitoring.

15 GEOTHERMAL ENERGY↗

Realistic galaxy image simulation via score-based generative models

ABSTRACT We show that a denoising diffusion probabilistic model (DDPM), a class of score-based generative model, can be used to produce realistic mock images that mimic observations of galaxies. Our method is tested with Dark Energy Spectroscopic Instrument (DESI) grz imaging of galaxies from the Photometry and Rotation curve OBservations from Extragalactic Surveys (PROBES) sample and galaxies selected from the Sloan Digital Sky Survey. Subjectively, the generated galaxies are highly realistic when compared with samples from the real data set. We quantify the similarity by borrowing from the deep generative learning literature, using the ‘Fréchet inception distance’ to test for subjective and morphological similarity. We also introduce the ‘synthetic galaxy distance’ metric to compare the emergent physical properties (such as total magnitude, colour, and half-light radius) of a ground truth parent and synthesized child data set. We argue that the DDPM approach produces sharper and more realistic images than other generative methods such as adversarial networks (with the downside of more costly inference), and could be used to produce large samples of synthetic observations tailored to a specific imaging survey. We demonstrate two potential uses of the DDPM: (1) accurate inpainting of occluded data, such as satellite trails, and (2) domain transfer, where new input images can be processed to mimic the properties of the DDPM training set. Here we ‘DESI-fy’ cartoon images as a proof of concept for domain transfer. Finally, we suggest potential applications for score-based approaches that could motivate further research on this topic within the astronomical community.

79 ASTRONOMY AND ASTROPHYSICS↗

Improving Limit Surface Search Algorithms in RAVEN Using Acceleration Schemes: Level II Milestone

The RAVEN code is becoming a comprehensive tool to perform Probabilistic Risk Assessment (PRA); Uncertainty Quantification (UQ) and Propagation; and Verification and Validation (V&V). The RAVEN code is being developed to support the Risk-Informed Safety Margin Characterization (RISMC) pathway by developing an advanced set of methodologies and algorithms for use in advanced risk analysis. The RISMC approach uses system simulator codes applied to stochastic analysis tools. The fundamental idea behind this coupling approach to perturb (by employing sampling strategies) timing and sequencing of events, internal parameters of the system codes (i.e., uncertain parameters of the physics model) and initial conditions to estimate values ranges and associated probabilities of figures of merit of interest for engineering and safety (e.g. core damage probability, etc.). This approach applied to complex systems such as nuclear power plants requires performing a series of computationally expensive simulation runs. The large computational burden is caused by the large set of (uncertain) parameters characterizing those systems. Consequently, exploring the uncertain/parametric domain, with a good level of confidence, is generally not affordable, considering the limited computational resources that are currently available. In addition, the recent tendency to develop newer tools, characterized by higher accuracy and larger computational resources (if compared with the presently used legacy codes, that have been developed decades ago), has made this issue even more compelling. In order to overcome to these limitations, the strategy for the exploration of the uncertain/parametric space needs to use at best the computational resources focusing the computational effort in those regions of the uncertain/parametric space that are “interesting” (e.g., risk-significant regions of the input space) with respect the targeted Figures Of Merit (FOM): for example, the failure of the system, subject of the analysis. These methodologies are named, in the RAVEN environment, adaptive sampling strategies. These methodologies infer system responses from surrogate models constructed from already existing samples (produced using high fidelity simulations) and suggest the most relevant location (coordinate in the input space) of the next sampling point to be explored in the uncertain/parametric domain. When using those methodologies, it is possible to understand features of the system response with a small number of carefully selected samples. This report focuses on the development and improvement of the limit surface search. The limit surface is an important concept in system reliability analysis. Without going into the details, which will be covered later in the report, the limit surface could be briefly described as an hyper-surface in the system uncertainty/parametric space separating the regions leading to a prescribed system outcome. For example, if the uncertainty/parametric space is the one generated by the reactor power level and the duration of the batteries, the system is a nuclear power plant and the system outcome discriminating variable is the clad failure in a station blackout scenario, then the limit surface separates the combinations of reactor power level and battery duration that lead to clad failure from the ones that do not.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗