Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “deep neural network”

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 523 records · Page 29
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Deep-learning-derived planetary boundary layer height from conventional meteorological measurements

Abstract. The planetary boundary layer (PBL) height (PBLH) is an important parameter for various meteorological and climate studies. This study presents a multi-structure deep neural network (DNN) model, which can estimate PBLH by integrating the morning temperature profiles and surface meteorological observations. The DNN model is developed by leveraging a rich dataset of PBLH derived from long-standing radiosonde records augmented with high-resolution micro-pulse lidar and Doppler lidar observations. We access the performance of the DNN with an ensemble of 10 members, each featuring distinct hidden-layer structures, which collectively yield a robust 27-year PBLH dataset over the southern Great Plains from 1994 to 2020. The influence of various meteorological factors on PBLH is rigorously analyzed through the importance test. Moreover, the DNN model's accuracy is evaluated against radiosonde observations and juxtaposed with conventional remote sensing methodologies, including Doppler lidar, ceilometer, Raman lidar, and micro-pulse lidar. The DNN model exhibits reliable performance across diverse conditions and demonstrates lower biases relative to remote sensing methods. In addition, the DNN model, originally trained over a plain region, demonstrates remarkable adaptability when applied to the heterogeneous terrains and climates encountered during the GoAmazon (Green Ocean Amazon; tropical rainforest) and CACTI (Cloud, Aerosol, and Complex Terrain Interactions; middle-latitude mountain) campaigns. These findings demonstrate the effectiveness of deep learning models in estimating PBLH, enhancing our understanding of boundary layer processes with implications for improving the representation of PBL in weather forecasting and climate modeling.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Physics-Informed Generative Adversarial Networks for Stochastic Differential Equations

We developed a new class of physics-informed generative adversarial networks (PI-GANs) to solve forward, inverse, and mixed stochastic problems in a unified manner based on a limited number of scattered measurements. Unlike standard GANs relying solely on data for training, here we encoded into the architecture of GANs the governing physical laws in the form of stochastic differential equations (SDEs) using automatic differentiation. In particular, we applied Wasserstein GANs with gradient penalty (WGAN-GP) for its enhanced stability compared to vanilla GANs. We first tested WGAN-GP in approximating Gaussian processes of different correlation lengths based on data realizations collected from simultaneous reads at sparsely placed sensors. We obtained good approximation of the generated stochastic processes to the target ones even if there is a mismatch between the input noise dimensionality and the effective dimensionality of the target stochastic processes. We also studied the overfitting issue for both the discriminator and the generator, and we found that overfitting occurs also in the generator in addition to the discriminator as previously reported. Subsequently, we considered the solution of elliptic SDEs requiring approximations of three stochastic processes, namely the solution, the forcing, and the diffusion coefficient. Here again, we assumed data collected from simultaneous reads at a limited number of sensors for the multiple stochastic processes. Furthermore, three generators were used for the PI-GANs: two of them were feed forward deep neural networks (DNNs), while the other one was the neural network induced by the SDE. For the case where we have one group of data, we employed one feed forward DNN as the discriminator, while for the case of multiple groups of data we employed multiple discriminators in PI-GANs. We solved forward, inverse, and mixed problems without changing the framework of PI-GANs, obtaining both the means and the standard deviations of the stochastic solution and the diffusion coefficient in good agreement with benchmarks. In this work, we have demonstrated the effectiveness of PI-GANs in solving SDEs for about 120 dimensions. In principle, PI-GANs could tackle very high dimensional problems given more sensor data with low-polynomial growth in computational cost.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Overcoming small minirhizotron datasets using transfer learning

Minirhizotron technology is widely used to study root growth and development. Yet, standard approaches for tracing roots in minirhiztron imagery is extremely tedious and time consuming. Machine learning approaches can help to automate this task. However, lack of enough annotated training data is a major limitation for the application of machine learning methods. Transfer learning is a useful technique to help with training when available datasets are limited. In this paper, we investigated the effect of pre-trained features from the massives-cale, irrelevant ImageNet dataset and a relatively moderate-scale, but relevant peanut root dataset on switchgrass root imagery segmentation applications. We compiled two minirhizotron image datasets to accomplish this study: one with 17,550 peanut root images and another with 28 switchgrass root images. Both datasets were paired with manually labeled ground truth masks. Deep neural networks based on the U-net architecture were used with different pre-trained features as initialization for automated, precise pixel-wise root segmentation in minirhizotron imagery. We observed that features pre-trained on a closely related but relatively moderate size dataset like our peanut dataset were more effective than features pre-trained on the large but unrelated ImageNet dataset. Here, we achieved high quality segmentation on peanut root dataset with 99.04% accuracy at the pixel-level and overcame errors in human-labeled ground truth masks. By applying transfer learning technique on limited switchgrass dataset with features pre-trained on peanut dataset, we obtained 99% segmentation accuracy in switchgrass imagery using only 21 images for training (fine tuning). Furthermore, the peanut pre-trained features can help the model converge faster and have much more stable performance.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

ASK: Adversarial Soft k-Nearest Neighbor Attack and Defense

K-Nearest Neighbor (kNN)-based deep learning methods have been applied to many applications due to their simplicity and geometric interpretability. However, the robustness of kNN-based deep classification models has not been thoroughly explored and kNN attack strategies are underdeveloped. In this paper, we first propose an Adversarial Soft kNN (ASK) loss for developing more effective kNN-based deep neural network attack strategies and designing better defense methods against them. Our ASK loss provides a differentiable surrogate of the expected kNN classification error. It is also interpretable as it preserves the mutual information between the perturbed input and the in-class-reference data. We use the ASK loss to design a novel attack method called the ASK-Attack (ASK-Atk), which shows superior attack efficiency and accuracy degradation relative to previous kNN attacks on hidden layers. We then derive an ASK-Defense (ASK-Def) method that optimizes the worst-case ASK training loss. Experiments on CIFAR-10 (ImageNet) show that (i) ASK-Atk achieves ≥13% (≥ 13% ) improvement in attack success rate over previous kNN attacks, and (ii) ASK-Def outperforms the conventional adversarial training method by ≥ 6.9% (≥ 3.5% ) in terms of robustness improvement. Relevant codes are available at https://github.com/wangren09/ASK .

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

DeepClone: Lightweight State Replication of Deep Learning Models for Data Parallel Training

Training modern deep neural network (DNN) models involves complex workflows triggered by model exploration, sensitivity analysis, explainability, etc. A key primitive in this context is the ability to clone a model training instance, i.e. “fork” the training process in a potentially different direction, which enables comparisons of different evolution paths using variations of training data and model parameters. However, in a quest improve the training throughput, a mix of data parallel, model parallel, pipeline parallel and layer-wise parallel approaches are making the problem of cloning highly complex. In this paper, we explore the problem of efficient cloning under such circumstances. To this end, we leverage several properties of data-parallel training and layer-wise parallelism to design DeepClone, a cloning approach based on augmenting the execution graph to gain direct access to tensors, which are then sharded and reconstructed asynchronously in order to minimize runtime overhead, standby duration, readiness duration. Compared with state-of-art approaches, DeepClone shows orders of magnitude improvement for several classes of DNN models.

data-parallel training↗
DOE OSTI2020

Reynolds-Averaged Turbulence Modeling Using Deep Learning with Local Flow Features: An Empirical Approach

Reynolds-Averaged Navier-Stoke (RANS) models offer an alternative avenue in predicting flow characteristics when the corresponding experiments are difficult to achieve due to geometry complexity, limited budget, or knowledge. RANS models require the knowledge of subgrid scale physics to solve conservation equations for mass, energy, and momentum. Mechanistic turbulence models, such as k-epsilon, are generally evaluated and calibrated for specific flow conditions with various degrees of uncertainty. These models have limited capability to assimilate a substantial amount of data due to model form constraints. Meanwhile, deep learning (DL) has been proven to be universal approximators with the potential to assimilate available, relevant, and adequately evaluated data. Moreover, deep neural networks (DNNs) can create surrogate models without knowing function forms. Such a data-driven approach can be used in updating fluid models based on observations as opposed to hard-wiring models with precalibrated correlations. The paper presents progress in applying DNNs to model Reynolds stress using two machine learning (ML) frameworks. A novel flow feature coverage mapping is proposed to quantify the physics coverage of DL-based closures. It can be used to examine the sufficiency of training data and input flow features for data-driven turbulence models. The case of a backward-facing step is formulated to demonstrate that not only can DNNs discover underlying correlation behind fluid data but also they can be implemented in RANS to predict flow characteristics without numerical stability issues. Finally, the presented research is a crucial stepping-stone toward the data-driven turbulence modeling, which potentially benefits the design of data-driven experiments that can be used to validate fluid models with ML-based fluid closures.

42 ENGINEERING↗
DOE OSTI2024

Anisotropic physics-regularized interpretable machine learning of microstructure evolution

Anisotropic Physics-Regularized Interpretable Machine Learning Microstructure Evolution (APRIMME) is a general-purpose machine learning solution for grain growth simulations. In prior work, PRIMME employed a deep neural network to predict site-specific migration as a function of its neighboring sites to model normal, isotropic, grain growth behavior. This work aims to extend this method by incorporating grain boundary misorientation-based grain growth behavior. APRIMME is trained on anisotropic simulations created using the Monte Carlo-Potts (MCP) model. Furthermore, the results of this work are compared statistically using grain radius, number of sides per grain, mean neighborhood misorientations, and the standard deviation of triple junction dihedral angles, and are found to match in most cases. The exceptions are small and seem to be related to two causes: (1) the deterministic model of APRIMME is learning from the stochastic simulations of MCP, which seems to accentuate triple junction behaviors; and, (2) a bias against very small grains is made evident in a quicker decrease in grains than expected at the beginning of an APRIMME simulation. APRIMME is also evaluated for its general ability to capture anisotropic grain growth behavior by first investigating different test case initial conditions, including a circle grain, three grain, and hexagonal grain microstructures.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Survey of Deep Learning and Physics-Based Approaches in Computational Wave Imaging

Computational wave imaging (CWI) extracts hidden structure and physical properties of a volume of material by analyzing wave signals that traverse that volume. Applications include seismic exploration of the Earth’s subsurface, acoustic imaging and nondestructive testing (NDT) in material science, and ultrasound computed tomography (USCT) in medicine. Current approaches for solving CWI problems can be divided into two categories: those rooted in traditional physics and those based on deep learning. Physics-based methods stand out for their ability to provide high-resolution and quantitatively accurate estimates of acoustic properties within the medium. However, they can be computationally intensive and are susceptible to ill-posedness and nonconvexity typical of CWI problems. Machine learning (ML)-based computational methods have recently emerged, offering a different perspective to address these challenges. Diverse scientific communities have independently pursued the integration of deep learning in CWI. This review discusses how contemporary scientific ML techniques, and deep neural networks in particular, have been developed to enhance and integrate with traditional physics-based methods for solving CWI problems. We present a structured framework that consolidates existing research spanning multiple domains, including computational imaging, wave physics, and data science. This study concludes with important lessons learned from existing ML-based methods and identifies technical hurdles and emerging trends through a systematic analysis of the extensive literature on this topic.

42 ENGINEERING↗
DOE OSTI2024

Bridging the Gap Between LLMs and LNS with Dynamic Data Format and Architecture Codesign

Deep neural networks (DNNs) have achieved tremendous success in the past few years. However, their training and inference demand exceptional computational and memory resources. Quantization has been shown as an effective approach to mitigate the cost, with the mainstream data types reduced from FP32 to FP16/BF16 and recently FP8 in the latest NVIDIA H100 GPUs. With increasingly aggressive quantization, however, the conventional floating-point formats suffer from limited precision in representing numbers around zero. Recently, NVIDIA demonstrated the potential of using a Logarithmic Number System (LNS) for the next generation of tensor cores. While LNS mitigates the hurdles in representing small numbers, in this work we observed a mismatch between LNS and the emerging Large Language Models (LLM), where LLM exhibits significant outliers when directly adopting the LNS format. In this paper, we present a data-format/architecture codesign to bright this gap. On the format side, we propose a dynamic LNS format to flexibly represent outliers at a higher precision, by exploiting asymmetry in the LNS representation and identifying outliers through a per-vector basis. On the architecture side, for demonstration, we realize the dynamic LNS format in a systolic array, which can handle the irregularity of the outliers at runtime. We implement our approach on an Alveo U280 FPGA as a prototype. Experimental results show that our design can effectively handle the outliers and resolve the mismatch between LNS and LLM, contributing to an accuracy improvement of 15.4% and 16% over the floating-point and the original LNS baselines, using four state-of-the-art LLM models. Our observation and design lay a solid foundation for the large-scale adoption of the LNS format in the next-generation deep learning hardware.

Haghi, Pouya↗
DOE OSTI2021

Progressive transfer learning for low-frequency data prediction in full waveform inversion

To effectively overcome the cycle-skipping issue in full waveform inversion (FWI), we developed a deep neural network (DNN) approach to predict the absent low-frequency components by exploiting the hidden physical relation connecting the low- and the high-frequency data. To efficiently solve this challenging nonlinear regression problem, two novel strategies were proposed to design the DNN architecture and to optimize the learning process: (1) dual data feed structure; (2) progressive transfer learning. With the dual data feed structure, not only the high-frequency data, but also the corresponding beat tone data are fed into the DNN to relieve the burden of feature extraction. The second strategy, progressive transfer learning, enables us to train the DNN using a single evolving training dataset. Within the framework of the progressive transfer learning, the training dataset continuously evolves in an iterative manner by gradually retrieving the subsurface information through the physics-based inversion module, progressively enhancing the prediction accuracy of the DNN and propelling the inversion process out of the local minima. Here, the synthetic numerical experiments suggest that, without any a priori geological information, the low-frequency data predicted by the progressive transfer learning are sufficiently accurate for an FWI engine to produce reliable subsurface velocity models free of cycle-skipping artifacts.

02 PETROLEUM↗
DOE OSTI2022

Large Scale Caching and Streaming of Training Data for Online Deep Learning

The training of deep neural network models on large data remains a difficult problem, despite progress towards scalable techniques. In particular, there is a mismatch between the random but predetermined order in which AI flows select training samples and the streaming I/O patterns for which traditional HPC data storage (e.g., parallel file systems) are designed. In addition, as more data are obtained, it is feasible neither simply to train learning models incrementally, due to catastrophic forgetting (i.e., bias towards new samples), nor to train frequently from scratch, due to prohibitive time and/or resource constraints. In this paper, we study data management techniques that combine caching and streaming with rehearsal support in order to enable efficient access to training samples in both offline training and continual learning. We revisit state-of-art streaming approaches based on data pipelines that transparently handle prefetching, caching, shuffling, and data augmentation, and discuss the challenges and opportunities that arise when combining these methods with data-parallel training techniques. We also report on preliminary experiments that evaluate the I/O overheads involved in accessing the training samples from a parallel file system (PFS) under several concurrency scenarios, highlighting the impact of the PFS on the design of the data pipelines.

data pipelines↗
DOE OSTI2021

An adaptive Hessian approximated stochastic gradient MCMC method

Bayesian approaches have been successfully integrated into training deep neural networks. One popular family is stochastic gradient Markov chain Monte Carlo methods (SG-MCMC), which have gained increasing interest due to their ability to handle large datasets and the potential to avoid overfitting. Although standard SG-MCMC methods have shown great performance in a variety of problems, they may be inefficient when the random variables in the target posterior densities have scale differences or are highly correlated. Here, we present an adaptive Hessian approximated stochastic gradient MCMC method to incorporate local geometric information while sampling from the posterior. The idea is to apply stochastic approximation (SA) to sequentially update a preconditioning matrix at each iteration. The preconditioner possesses second-order information and can guide the random walk of a sampler efficiently. Instead of computing and saving the full Hessian of the log posterior, we use limited memory of the samples and their stochastic gradients to approximate the inverse Hessian-vector multiplication in the updating formula. Moreover, by smoothly optimizing the preconditioning matrix via SA, our proposed algorithm can asymptotically converge to the target distribution with a controllable bias under mild conditions. To reduce the training and testing computational burden, we adopt a magnitude-based weight pruning method to enforce the sparsity of the network. Our method is user-friendly and demonstrates better learning results compared to standard SG-MCMC updating rules. The approximation of inverse Hessian alleviates storage and computational complexities for large dimensional models. Numerical experiments are performed on several problems, including sampling from 2D correlated distribution, synthetic regression problems, and learning the numerical solutions of heterogeneous elliptic PDE. The numerical results demonstrate great improvement in both the convergence rate and accuracy.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Filament-Free Bulk Resistive Memory Enables Deterministic Analogue Switching

Digital computing is nearing its physical limits as computing needs and energy consumption rapidly increase. Analogue-memory-based neuromorphic computing can be orders of magnitude more energy efficient at data-intensive tasks like deep neural networks, but has been limited by the inaccurate and unpredictable switching of analogue resistive memory. Filamentary resistive random access memory (RRAM) suffers from stochastic switching due to the random kinetic motion of discrete defects in the nanometer-sized filament. Here, this stochasticity is overcome by incorporating a solid electrolyte interlayer, in this case, yttria-stabilized zirconia (YSZ), toward eliminating filaments. Filament-free, bulk-RRAM cells instead store analogue states using the bulk point defect concentration, yielding predictable switching because the statistical ensemble behavior of oxygen vacancy defects is deterministic even when individual defects are stochastic. Both experiments and modeling show bulk-RRAM devices using TiO2-X switching layers and YSZ electrolytes yield deterministic and linear analogue switching for efficient inference and training. Bulk-RRAM solves many outstanding issues with memristor unpredictability that have inhibited commercialization, and can, therefore, enable unprecedented new applications for energy-efficient neuromorphic computing. Beyond RRAM, this work shows how harnessing bulk point defects in ionic materials can be used to engineer deterministic nanoelectronic materials and devices.

36 MATERIALS SCIENCE↗
DOE OSTI2025

AI‐Driven Defect Engineering for Advanced Thermoelectric Materials

Thermoelectric materials offer a promising pathway to directly convert waste heat to electricity. However, achieving high performance remains challenging due to intrinsic trade-offs between electrical conductivity, the Seebeck coefficient, and thermal conductivity, which are further complicated by the presence of defects. This review explores how artificial intelligence (AI) and machine learning (ML) are transforming thermoelectric materials design. Advanced ML approaches including deep neural networks, graph-based models, and transformer architectures, integrated with high-throughput simulations and growing databases, effectively capture structure-property relationships in a complex multiscale defect space and overcome the “curse of dimensionality”. This review discusses AI-enhanced defect engineering strategies such as composition optimization, entropy and dislocation engineering, and grain boundary design, along with emerging inverse design techniques for generating materials with targeted properties. Finally, it outlines future opportunities in novel physics mechanisms and sustainability, highlighting the critical role of AI in accelerating the discovery of thermoelectric materials.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Nonvolatile Electrochemical Random‐Access Memory under Short Circuit

Abstract Electrochemical random‐access memory (ECRAM) is a recently developed and highly promising analog resistive memory element for in‐memory computing. One longstanding challenge of ECRAM is attaining retention time beyond a few hours. This short retention has precluded ECRAM from being considered for inference classification in deep neural networks, which is likely the largest opportunity for in‐memory computing. In this work, an ECRAM cell with orders of magnitude longer retention than previously achieved is developed, and which is anticipated to exceed ten years at 85 °C. This study hypothesizes that the origin of this exceptional retention is phase separation, which enables the formation of multiple effectively equilibrium resistance states. This work highlights the promises and opportunities to use phase separation to yield ECRAM cells with exceptionally long, and potentially permanent, retention times.

Kim, Diana S.↗
DOE OSTI2025

Effect of Ambient Organic Acids on the Water Structure at ${\rm TiO}_{2}$ Interfaces

A molecular-level understanding of the effects of ambient organic compounds on the wettability of titanium dioxide ( ${\rm TiO}_{2}$ ) surfaces is relevant to many of its energy-related and environmental applications. Herein, we focus on two common atmospheric carboxylic acids, formic and acetic acid, and characterize their adsorption/ desorption at the aqueous interfaces of anatase and rutile ${\rm TiO}_{2}$ using molecular dynamics with an ab initio deep neural network potential. Our simulations show that these acids prefer to be localized in the interfacial water layers close to the ${\rm TiO}_{2}$ surface where they are stabilized by the interaction/exchange of their acid proton with a surface oxygen, rather than chemisorb at the surface Ti sites by displacing the adsorbed water. Notably, these acids make the surface of anatase hydrophobic, whereas the larger fraction of adsorbed water dissociation can offset their effect on rutile. Furthermore, these results provide a picture where carboxylic acids control the wettability of ${\rm TiO}_{2}$ largely through acid-base chemistry at the interface rather than chemisorption on the oxide surface, a finding that can help improve the design of self-cleaning surfaces and photocatalytic devices.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Toward a Generalizable Prediction Model of Molten Salt Mixture Density with Chemistry‐Informed Transfer Learning

Optimally designing applications of molten salts requires knowledge of their thermophysical properties over a wide range of temperatures and compositions. There exist significant gaps in existing databases and this data can be challenging to experimentally measure due to high temperatures, salt corrosivity, and salt hygroscopicity. Existing databases have been used to create Redlich–Kister (RK) models for mixture density showing improved accuracy with respect to ideal mixing assumptions, but these models require subcomponent data measurements for each new system, therefore lacking generality. In order to address generalizability and data sparsity, a transfer learning procedure is proposed to train deep neural networks (DNNs) using a combination of semi‐empirical relationships (RK), data from the thermophysical arm of the molten salt thermal properties database and universal ab initio properties of component mixtures taken from the joint automated repository for various integrated simulations (JARVIS) classical force‐field inspired descriptors database to predict density in molten salts. Herein, it is shown that DNNs predict molten salt density with an r 2 over 0.99 and a mean absolute percentage error under 1%, outperforming alternative methods.

inorganic materials↗