Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Deep Neural Networks”

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 181 records · Page 10

Measuring the Energy Consumption and Efficiency of Deep Neural Networks: An Empirical Analysis and Design Recommendations

Addressing the "Red-AI" trend of rising energy consumption by large-scale neural networks, this study investigates the measured energy consumption of training various fully connected neural network architectures. We introduce the BUTTER-E dataset, an augmentation to the BUTTER Empirical Deep Learning dataset, containing energy consumption and performance data from 41,129 individual experimental runs spanning 30,582 distinct configurations: 13 datasets, 20 sizes (trainable parameters), 8 "shapes", and 14 depths on both CPUs and GPUs using node-level watt-meters. This dataset reveals the complex relationship between dataset size, network structure, and energy use. Our analysis uncovers a surprising, hardware-mediated non-linear relationship between energy efficiency and network design, challenging the assumption that reducing the number of parameters or FLOPs is the best way to achieve greater energy efficiency. We propose a straightforward and effective energy model that accounts for network size, computing, and memory hierarchy. Highlighting the need for cache-considerate algorithm development, we suggest a codesign approach to energy efficient network, algorithm, and hardware design. This work contributes to the fields of sustainable computing and Green AI, offering practical guidance for creating more energy-efficient neural networks and promoting sustainable AI.

97 MATHEMATICS AND COMPUTING↗

Prediction of inter-chain distance maps of protein complexes with 2D attention-based deep neural networks

Residue-residue distance information is useful for predicting tertiary structures of protein monomers or quaternary structures of protein complexes. Many deep learning methods have been developed to predict intra-chain residue-residue distances of monomers accurately, but few methods can accurately predict inter-chain residue-residue distances of complexes. We develop a deep learning method CDPred (i.e., Complex Distance Prediction) based on the 2D attention-powered residual network to address the gap. Tested on two homodimer datasets, CDPred achieves the precision of 60.94% and 42.93% for top L/5 inter-chain contact predictions (L: length of the monomer in homodimer), respectively, substantially higher than DeepHomo’s 37.40% and 23.08% and GLINTER’s 48.09% and 36.74%. Tested on the two heterodimer datasets, the top Ls/5 inter-chain contact prediction precision (Ls: length of the shorter monomer in heterodimer) of CDPred is 47.59% and 22.87% respectively, surpassing GLINTER’s 23.24% and 13.49%. Moreover, the prediction of CDPred is complementary with that of AlphaFold2-multimer.

59 BASIC BIOLOGICAL SCIENCES↗

Multilevel-in-width training for deep neural network regression

A common challenge in regression is that for many problems, the degrees of freedom required for a high-quality solution also allows for overfitting. Regularization is a class of strategies that seek to restrict the range of possible solutions so as to discourage overfitting while still enabling good solutions, and different regularization strategies impose different types of restrictions. In this paper, we present a multilevel regularization strategy that constructs and trains a hierarchy of neural networks, each of which has layers that are wider versions of the previous network's layers. Here, we draw intuition and techniques from the field of Algebraic Multigrid (AMG), traditionally used for solving linear and nonlinear systems of equations, and specifically adapt the Full Approximation Scheme (FAS) for nonlinear systems of equations to the problem of deep learning. Training through V-cycles then encourage the neural networks to build a hierarchical understanding of the problem. We refer to this approach as multilevel-in-width to distinguish from prior multilevel works which hierarchically alter the depth of neural networks. The resulting approach is a highly flexible framework that can be applied to a variety of layer types, which we demonstrate with both fully connected and convolutional layers. We experimentally show with PDE regression problems that our multilevel training approach is an effective regularizer, improving the generalize performance of the neural networks studied.

97 MATHEMATICS AND COMPUTING↗

Solar Transposition Modeling via Deep Neural Networks With Sky Images

This article presents a notable advance toward the development of a new method of increasing the single-axis tracking photovoltaic (PV) system power output by improving the determination and near-term prediction of the optimum module tilt angle. The tilt angle of the plane receiving the greatest total irradiance changes with Sun position and atmospheric conditions including cloud formation and movement, aerosols, and particulate loading, as well as varying albedo within a module's field of view. In this article, we present a multi-input convolutional neural network that can create a profile of plane-of-array irradiance versus surface tilt angle over a full 180° arc from horizon to horizon. As input, the neural network uses the calculated solar position and clear-sky irradiance values, along with sky images. The target irradiance values are provided by the multiplanar irradiance sensor (MPIS). In order to account for varying irradiance conditions, the MPIS signal is normalized by the theoretical clear-sky global horizontal irradiance. Using this information, the neural network outputs an N -dimensional vector, where N is the number of points to approximate the MPIS curve via Fourier resampling. The output vector of the model is smoothed with a Gaussian kernel to account for error in the downsamping and subsequent upsampling steps, as well as to smooth the unconstrained output of the model. Furthermore, these profiles may be used to perform near-term prediction of angular irradiance, which can then inform the movement of a PV tracker.

42 ENGINEERING↗

SuperNeurons: Dynamic GPU Memory Management for Training Deep Neural Networks

Going deeper and wider in neural architectures improves their accuracy, while the limited GPU DRAM places an undesired restriction on the network design domain. Deep Learning (DL) practitioners either need to change to less de- sired network architectures, or nontrivially dissect a network across multiGPUs. These distract DL practitioners from concentrating on their original machine learning tasks. We present SuperNeurons: a dynamic GPU memory scheduling runtime to enable the network training far beyond the GPU DRAM capacity. SuperNeurons features 3 memory optimizations, Liveness Analysis, Unified Tensor Pool, and Cost-Aware Recomputation; together they effectively reduce the network-wide peak memory usage down to the maximal memory usage among layers. We also address the performance issues in these memory-saving techniques. Given the limited GPU DRAM, SuperNeurons not only provisions the necessary memory for the training, but also dynamically allocates the memory for convolution workspaces to achieve the high performance. Evaluations against Caffe, Torch, MXNet and TensorFlow have demonstrated that SuperNeurons trains at least 3.2432 deeper network than current ones with the leading performance. Particularly, SuperNeurons can train ResNet2500 that has 104 basic network layers on a 12GB K40c.

Wang, Linnan↗

Peculiar velocity estimation from kinetic SZ effect using deep neural networks

ABSTRACT The Sunyaev–Zel’dolvich (SZ) effect is expected to be instrumental in measuring velocities of distant clusters in near future telescope surveys. We simplify the calculation of peculiar velocities of galaxy clusters using deep learning frameworks trained on numerical simulations to avoid the independent estimation of the optical depth. Images of distorted photon backgrounds are generated for idealized observations using one of the largest cosmological hydrodynamical simulations, the Magneticum simulations. The model is tested to determine its ability of estimating peculiar velocities from future kinetic SZ observations under different noise conditions. The deep learning algorithm displays robustness in estimating peculiar velocities from kinetic SZ effect by an improvement in accuracy of about 17 per cent compared to the analytical approach.

Wang, Yuyu (ORCID:0000000202458547)↗

Characterizing Structure-Dependent TiS 2 /Water Interfaces Using Deep-Neural-Network-Assisted Molecular Dynamics

As a promising layered electrode material, TiS 2 -based capacitive deionization (CDI) devices for water desalination have attracted significant attention. However, TiS 2 /H 2 O interfacial features, potentially important for device optimization, remain unidentified. Using Deep Potential Molecular Dynamics (DPMD), we characterized distinct aqueous interfaces introduced by four TiS 2 terminations expected to be present as water intercalates into TiS 2 , namely, Armchair, Zigzag, Zigzag-L, and Zigzag-R. First, we assessed important representative physical properties of the system to validate the deep potentials (DPs). DPMD simulations agree well with experiments and first-principles simulations, suggesting the DPs are accurate and reliable. Subsequent simulations of these TiS 2 /water interfaces revealed how TiS 2 surface termination influences the structure of interfacial water. This effect is most evident in the first and second water layers close to the TiS 2 surface, and more pronounced when spontaneous dissociative adsorption of water occurs. The extent of water dissociation on each surface was evaluated using enhanced sampling. Zigzag-L is the only interface where proton transfer from adsorbed water to TiS 2 surface S atoms is thermodynamically and kinetically favored. The coexistence of surface four-fold-coordinated Ti (Ti 4c ) and one-fold-coordinated S (S 1c ) is found to be essential to making proton transfer feasible on the Zigzag-L surface. Furthermore, remaining unprotonated S 1c atoms can act as good proton acceptors after water dissociation. Thus, TiS 2 with Zigzag-L termination may be a surface to avoid in CDI device construction, given that pH fluctuations adversely affect performance. Furthermore, this work provides new understanding of TiS 2 /H 2 O interfacial features that could aid future design and optimization of TiS 2 -based CDI devices for water desalination.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

KamNet: An integrated spatiotemporal deep neural network for rare event searches in KamLAND-Zen

Rare event searches allow us to search for new physics at energy scales inaccessible with other means by leveraging specialized large-mass detectors. Machine learning provides a new tool to maximize the information provided by these detectors. The information is sparse, which forces these algorithms to start from the lowest level data and exploit all symmetries in the detector to produce results. In this work we present KamNet, which harnesses breakthroughs in geometric deep learning and spatiotemporal data analysis to maximize the physics reach of KamLAND-Zen, a kiloton scale spherical liquid scintillator detector searching for 0⁢νβ⁢β. Using a simplified background model for KamLAND, we show that KamNet outperforms a conventional convolutional neural network (CNN) on benchmarking Monte Carlo simulations with an increasing level of robustness. Using simulated data, we then demonstrate KamNet's ability to increase KamLAND-Zen's sensitivity to 0ν⁢ββ and 2⁢ν⁢β⁢β decay to excited states. A key component of this work is the addition of an attention mechanism to elucidate the underlying physics KamNet is using for the background rejection.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DeepGRN: prediction of transcription factor binding site across cell-types using attention-based deep neural networks

Abstract Background Due to the complexity of the biological systems, the prediction of the potential DNA binding sites for transcription factors remains a difficult problem in computational biology. Genomic DNA sequences and experimental results from parallel sequencing provide available information about the affinity and accessibility of genome and are commonly used features in binding sites prediction. The attention mechanism in deep learning has shown its capability to learn long-range dependencies from sequential data, such as sentences and voices. Until now, no study has applied this approach in binding site inference from massively parallel sequencing data. The successful applications of attention mechanism in similar input contexts motivate us to build and test new methods that can accurately determine the binding sites of transcription factors. Results In this study, we propose a novel tool (named DeepGRN) for transcription factors binding site prediction based on the combination of two components: single attention module and pairwise attention module. The performance of our methods is evaluated on the ENCODE-DREAM in vivo Transcription Factor Binding Site Prediction Challenge datasets. The results show that DeepGRN achieves higher unified scores in 6 of 13 targets than any of the top four methods in the DREAM challenge. We also demonstrate that the attention weights learned by the model are correlated with potential informative inputs, such as DNase-Seq coverage and motifs, which provide possible explanations for the predictive improvements in DeepGRN. Conclusions DeepGRN can automatically and effectively predict transcription factor binding sites from DNA sequences and DNase-Seq coverage. Furthermore, the visualization techniques we developed for the attention modules help to interpret how critical patterns from different types of input features are recognized by our model.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Detecting rare neutral atomic-carbon absorbers with a deep neural network

ABSTRACT C i absorbers play an important role as indicators for exploring the presence of cold gas in the interstellar medium of galaxies. However, the current data base of C i absorbers is very limited due to their weak absorption feature and rarity. Here, we report results from a search of C i λλ1560, 1656 absorption lines using Mg ii absorbers as signposts with modified deep learning algorithms, which provides a very quick way to search for weak C i absorber candidates. A total of 107 C i absorbers were detected, which nearly doubles the size of previously known samples. In addition, we found 17 C i absorbers to be associated with 2175 Å dust absorbers (2DAs), i.e. about 16 per cent C i absorbers are associated with 2DAs. Comparing the average dust depletion patterns of C i absorbers with those of damped Lyman α absorbers (DLAs), Mg ii absorbers, Ca ii absorbers, and 2175 Å dust absorbers (2DAs) shows that C i absorbers generally have environments with more dust than DLAs, Mg ii, and Ca ii absorbers, but similar to dust in 2DAs. Similarity between the dust depletion pattern of C i absorbers to that of the warm disc in the Milky Way indicates that C i absorption clouds are possibly associated with disc components in distant galaxies. Therefore, C i absorbers are confirmed to be excellent probes to trace cold gas and dust in the Universe.

Ge, Jian↗

Sparse coding of pathology slides compared to transfer learning with deep neural networks

Background Histopathology images of tumor biopsies present unique challenges for applying machine learning to the diagnosis and treatment of cancer. The pathology slides are high resolution, often exceeding 1GB, have non-uniform dimensions, and often contain multiple tissue slices of varying sizes surrounded by large empty regions. The locations of abnormal or cancerous cells, which may constitute a small portion of any given tissue sample, are not annotated. Cancer image datasets are also extremely imbalanced, with most slides being associated with relatively common cancers. Since deep representations trained on natural photographs are unlikely to be optimal for classifying pathology slide images, which have different spectral ranges and spatial structure, we here describe an approach for learning features and inferring representations of cancer pathology slides based on sparse coding. Results We show that conventional transfer learning using a state-of-the-art deep learning architecture pre-trained on ImageNet (RESNET) and fine tuned for a binary tumor/no-tumor classification task achieved between 85% and 86% accuracy. However, when all layers up to the last convolutional layer in RESNET are replaced with a single feature map inferred via a sparse coding using a dictionary optimized for sparse reconstruction of unlabeled pathology slides, classification performance improves to over 93%, corresponding to a 54% error reduction. Conclusions We conclude that a feature dictionary optimized for biomedical imagery may in general support better classification performance than does conventional transfer learning using a dictionary pre-trained on natural images.

59 BASIC BIOLOGICAL SCIENCES↗

Cosmic Ray Background Removal With Deep Neural Networks in SBND

In liquid argon time projection chambers exposed to neutrino beams and running on or near surface levels, cosmic muons, and other cosmic particles are incident on the detectors while a single neutrino-induced event is being recorded. In practice, this means that data from surface liquid argon time projection chambers will be dominated by cosmic particles, both as a source of event triggers and as the majority of the particle count in true neutrino-triggered events. In this work, we demonstrate a novel application of deep learning techniques to remove these background particles by applying deep learning on full detector images from the SBND detector, the near detector in the Fermilab Short-Baseline Neutrino Program. We use this technique to identify, on a pixel-by-pixel level, whether recorded activity originated from cosmic particles or neutrino interactions.

79 ASTRONOMY AND ASTROPHYSICS↗

Million-scale data integrated deep neural network for phonon properties of heuslers spanning the periodic table

Existing machine learning potentials for predicting phonon properties of crystals are typically limited on a material-to-material basis, primarily due to the exponential scaling of model complexity with the number of atomic species. We address this bottleneck with the developed Elemental Spatial Density Neural Network Force Field, namely Elemental-SDNNFF. The effectiveness and precision of our Elemental-SDNNFF approach are demonstrated on 11,866 full, half, and quaternary Heusler structures spanning 55 elements in the periodic table by prediction of complete phonon properties. Self-improvement schemes including active learning and data augmentation techniques provide an abundant 9.4 million atomic data for training. Deep insight into predicted ultralow lattice thermal conductivity (<1 Wm –1 K –1 ) of 774 Heusler structures is gained by p–d orbital hybridization analysis. Additionally, a class of two-band charge-2 Weyl points, referred to as “double Weyl points”, are found in 68% and 87% of 1662 half and 1550 quaternary Heuslers, respectively.

36 MATERIALS SCIENCE↗

Discovery of widespread transcription initiation at microsatellites predictable by sequence-based deep neural network

Using the Cap Analysis of Gene Expression (CAGE) technology, the FANTOM5 consortium provided one of the most comprehensive maps of transcription start sites (TSSs) in several species. Strikingly, ~72% of them could not be assigned to a specific gene and initiate at unconventional regions, outside promoters or enhancers. Here, we probe these unassigned TSSs and show that, in all species studied, a significant fraction of CAGE peaks initiate at microsatellites, also called short tandem repeats (STRs). To confirm this transcription, we develop Cap Trap RNA-seq, a technology which combines cap trapping and long read MinION sequencing. We train sequence-based deep learning models able to predict CAGE signal at STRs with high accuracy. These models unveil the importance of STR surrounding sequences not only to distinguish STR classes, but also to predict the level of transcription initiation. Importantly, genetic variants linked to human diseases are preferentially found at STRs with high transcription initiation level, supporting the biological and clinical relevance of transcription initiation at STRs. Together, our results extend the repertoire of non-coding transcription associated with DNA tandem repeats and complexify STR polymorphism.

59 BASIC BIOLOGICAL SCIENCES↗

Learning Deep Neural Network Representations for Koopman Operators of Nonlinear Dynamical Systems

The Koopman operator has recently garnered much attention for its value in dynamical systems analysis and data-driven model discovery. However, its application has been hindered by the computational complexity of extended dynamic mode decomposition; this requires a combinatorially large basis set to adequately describe many nonlinear systems of interest, e.g. cyber-physical infrastructure systems, biological networks, social systems, and fluid dynamics. Often the dictionaries generated for these problems are manually curated, requiring domain-specific knowledge and painstaking tuning. In this paper we introduce a deep learning framework for learning Koopman operators of nonlinear dynamical systems. We show that this novel method automatically selects efficient deep dictionaries, outperforming state-of-the-art methods. We benchmark this method on partially observed nonlinear systems, including the glycolytic oscillator and show it is able to predict quantitatively 100 steps into the future, using only a single timepoint, and qualitative oscillatory behavior 400 steps into the future.

koopman operators, machine learning, deep learning↗

A fine pore-preserved deep neural network for porosity analytics of a high burnup U-10Zr metallic fuel

Abstract U-10 wt.% Zr (U-10Zr) metallic fuel is the leading candidate for next-generation sodium-cooled fast reactors. Porosity is one of the most important factors that impacts the performance of U-10Zr metallic fuel. The pores generated by the fission gas accumulation can lead to changes in thermal conductivity, fuel swelling, Fuel-Cladding Chemical Interaction (FCCI) and Fuel-Cladding Mechanical Interaction (FCMI). Therefore, it is crucial to accurately segment and analyze porosity to understand the U-10Zr fuel system to design future fast reactors. To address the above issues, we introduce a workflow to process and analyze multi-source Scanning Electron Microscope (SEM) image data. Moreover, an encoder-decoder-based, deep fully convolutional network is proposed to segment pores accurately by integrating the residual unit and the densely-connected units. Two SEM 250 × field of view image datasets with different formats are utilized to evaluate the new proposed model’s performance. Sufficient comparison results demonstrate that our method quantitatively outperforms two popular deep fully convolutional networks. Furthermore, we conducted experiments on the third SEM 2500 × field of view image dataset, and the transfer learning results show the potential capability to transfer the knowledge from low-magnification images to high-magnification images. Finally, we use a pre-trained network to predict the pores of SEM images in the whole cross-sectional image and obtain quantitative porosity analysis. Our findings will guide the SEM microscopy data collection efficiently, provide a mechanistic understanding of the U-10Zr fuel system and bridge the gap between advanced characterization to fuel system design.

36 MATERIALS SCIENCE↗

Recovering missing CFD data for high-order discretizations using deep neural networks and dynamics learning

Data I/O poses a significant bottleneck in large-scale CFD simulations; thus, practitioners would like to significantly reduce the number of times the solution is saved to disk, yet retain the ability to recover any field quantity (at any time instance) a posteriori . The objective of this work is therefore to accurately recover missing CFD data a posteriori at any time instance, given that the solution has been written to disk at only a relatively small number of time instances. We consider in particular high-order discretizations (e.g., discontinuous Galerkin), as such techniques are becoming increasingly popular for the simulation of highly separated flows. To satisfy this objective, this work proposes a methodology consisting of two stages: 1) dimensionality reduction and 2) dynamics learning. For dimensionality reduction, we propose a novel hierarchical approach. First, the method reduces the number of degrees of freedom within each element of the high-order discretization by applying autoencoders from deep learning. Second, the methodology applies principal component analysis to compress the global vector of encodings. This leads to a low-dimensional state, which associates with a nonlinear embedding of the original CFD data. For dynamics learning, we propose to apply regression techniques (e.g., kernel methods) to learn the discrete-time velocity characterizing the time evolution of this low-dimensional state. In conclusion, a numerical example on a large-scale CFD example characterized by nearly 13 million degrees of freedom illustrates the suitability of the proposed method in an industrial setting.

97 MATHEMATICS AND COMPUTING↗

Quantifying leaf symptoms of sorghum charcoal rot in images of field‐grown plants using deep neural networks

Abstract Charcoal rot of sorghum (CRS) is a significant disease affecting sorghum crops, with limited genetic resistance available. The causative agent, Macrophomina phaseolina (Tassi) Goid, is a highly destructive fungal pathogen that targets over 500 plant species globally, including essential staple crops. Utilizing field image data for precise detection and quantification of CRS could greatly assist in the prompt identification and management of affected fields and thereby reduce yield losses. The objective of this work was to implement various machine learning algorithms to evaluate their ability to accurately detect and quantify CRS in red‐green‐blue images of sorghum plants exhibiting symptoms of infection. EfficientNet‐B3 and a fully convolutional network emerged as the top‐performing models for image classification and segmentation tasks, respectively. Among the classification models evaluated, EfficientNet‐B3 demonstrated superior performance, achieving an accuracy of 86.97%, a recall rate of 0.71, and an F1 score of 0.73. Of the segmentation models tested, FCN proved to be the most effective, exhibiting a validation accuracy of 97.76%, a recall rate of 0.68, and an F1 score of 0.66. As the size of the image patches increased, both models’ validation scores increased linearly, and their inference time decreased exponentially. This trend could be attributed to larger patches containing more information, improving model performance, and fewer patches reducing the computational load, thus decreasing inference time. The models, in addition to being immediately useful for breeders and growers of sorghum, advance the domain of automated plant phenotyping and may serve as a foundation for drone‐based or other automated field phenotyping efforts. Additionally, the models presented herein can be accessed through a web‐based application where users can easily analyze their own images.

Gonzalez, Emmanuel M.↗