Search NASA⌕ Search

SEARCH · Search NASA

Results for “deep learning method”

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 1,171 records · Page 65

Background rejection in NEXT using deep neural networks

Here, we investigate the potential of using deep learning techniques to reject background events in searches for neutrinoless double beta decay with high pressure xenon time projection chambers capable of detailed track reconstruction. The differences in the topological signatures of background and signal events can be learned by deep neural networks via training over many thousands of events. These networks can then be used to classify further events as signal or background, providing an additional background rejection factor at an acceptable loss of efficiency. The networks trained in this study performed better than previous methods developed based on the use of the same topological signatures by a factor of 1.2 to 1.6, and there is potential for further improvement.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

The PAU Survey: Photometric redshifts using transfer learning from simulations

In this paper, we introduce the DEEPZ deep learning photometric redshift (photo-z) code. As a test case, we apply the code to the PAU survey (PAUS) data in the COSMOS field. DEEPZ reduces the σ68 scatter statistic by 50 percent at iAB = 22.5 compared to existing algorithms. This improvement is achieved through various methods, including transfer learning from simulations where the training set consists of simulations as well as observations, which reduces the need for training data. The redshift probability distribution is estimated with a mixture density network (MDN), which produces accurate redshift distributions. Our code includes an autoencoder to reduce noise and extract features from the galaxy SEDs. It also benefits from combining multiple networks, which lowers the photo-z scatter by 10 percent. Furthermore, training with randomly constructed coadded fluxes adds information about individual exposures, reducing the impact of photometric outliers. In addition to opening up the route for higher redshift precision with narrow bands, these machine learning techniques can also be valuable for broad-band surveys.

79 ASTRONOMY AND ASTROPHYSICS↗

REV-INR: Regularized Evidential Implicit Neural Representation for Uncertainty-Aware Volume Visualization

Applications of Implicit Neural Representations (INRs) have emerged as a promising deep learning approach for compactly representing large volumetric datasets. These models can act as surrogates for volume data, enabling efficient storage and on-demand reconstruction via model predictions. However, conventional deterministic INRs only provide value predictions without insights into the model’s prediction uncertainty or the impact of inherent noisiness in the data. This limitation can lead to unreliable data interpretation and visualization due to prediction inaccuracies in the reconstructed volume. Identifying erroneous results extracted from model-predicted data may be infeasible, as raw data may be unavailable due to its large size. To address this challenge, we introduce REV-INR, Regularized Evidential Implicit Neural Representation, which learns to predict data values accurately along with the associated coordinate-level data uncertainty and model uncertainty using only a single forward pass of the trained REV-INR during inference. By comprehensively comparing and contrasting REV-INR with existing well-established deep uncertainty estimation methods, we show that REV-INR achieves the best volume reconstruction quality with robust data (aleatoric) and model (epistemic) uncertainty estimates using the fastest inference time. Consequently, we demonstrate that REV-INR facilitates assessment of the reliability and trustworthiness of the extracted isosurfaces and volume visualization results, enabling analyses to be solely driven by model-predicted data.

Saklani, Shanu [Indian Institute of Technology, Ka↗

Machine Learning for Turbulence Modeling

This work was conducted as part of a Harry S. Truman Fellowship Laboratory Directed Research and Development project. The goal was to use machine learning methods to provide uncertainty quantification and model improvements for Reynolds Averaged Navier Stokes (RANS) turbulence models. For applications of interest in energy, safety, and security, it is critical to be able to model turbulence accurately. Current RANS models are unreliable for many flows of engineering relevance. Machine learning provides an avenue for developing improved models based on the data generated by high fidelity simulations. In this project, machine learning methods were used to predict when current RANS models would fail. They were also used to develop improved RANS closure models. A key aim was developing a tight feedback loop between scientific domain knowledge and data driven methods. To this end, a methodology for incorporating known invariance constraints into the machine learning models was proposed and evaluated. This work demonstrated that incorporating known constraints into the data driven models provided improved performance and reduced computational cost. This research represents one of the first applications of deep learning to turbulence modeling.

42 ENGINEERING↗

Machine learning models for PDE constrained optimization

Partial differential equation (PDE)-constrained optimization problems arise in a variety of scientific and engineering applications, such as topology optimization, electrodynamics, fluid dynamics, and structural dynamics. However, these problems are often challenging and computationally expensive to solve, due to the need to solve the PDEs within the optimization loop. One approach to reducing the computational cost of these methods while providing convergence guarantees is through inexact trust region methods; this method uses lower fidelity solutions of the PDE at early stages of the optimization and adjusts the required accuracy of inexact PDE solvers as the optimization progresses. In this work, we explore the use of machine learning based surrogate models with these inexact trust region methods. We first demonstrate the potential of this approach by using Gaussian processes as the surrogate model and test this on a simple PDE-constrained optimization problem. We then document explorations into improving the computational costs of evolutional deep neural network / neural Galerkin methods, with the eventual goal of using these methods with the inexact trust region algorithms. We are able to speed up these approaches, albeit at the cost of lower accuracy.

97 MATHEMATICS AND COMPUTING↗

Physics-Guided, Physics-Informed, and Physics-Encoded Neural Networks and Operators in Scientific Computing: Fluid and Solid Mechanics

Abstract Advancements in computing power have recently made it possible to utilize machine learning and deep learning to push scientific computing forward in a range of disciplines, such as fluid mechanics, solid mechanics, materials science, etc. The incorporation of neural networks is particularly crucial in this hybridization process. Due to their intrinsic architecture, conventional neural networks cannot be successfully trained and scoped when data are sparse, which is the case in many scientific and engineering domains. Nonetheless, neural networks provide a solid foundation to respect physics-driven or knowledge-based constraints during training. Generally speaking, there are three distinct neural network frameworks to enforce the underlying physics: (i) physics-guided neural networks (PgNNs), (ii) physics-informed neural networks (PiNNs), and (iii) physics-encoded neural networks (PeNNs). These methods provide distinct advantages for accelerating the numerical modeling of complex multiscale multiphysics phenomena. In addition, the recent developments in neural operators (NOs) add another dimension to these new simulation paradigms, especially when the real-time prediction of complex multiphysics systems is required. All these models also come with their own unique drawbacks and limitations that call for further fundamental research. This study aims to present a review of the four neural network frameworks (i.e., PgNNs, PiNNs, PeNNs, and NOs) used in scientific computing research. The state-of-the-art architectures and their applications are reviewed, limitations are discussed, and future research opportunities are presented in terms of improving algorithms, considering causalities, expanding applications, and coupling scientific and deep learning solvers.

Computer Science↗

Deep learning: Extrapolation tool for ab initio nuclear theory

Ab initio approaches in nuclear theory, such as the no-core shell model (NCSM), have been developed for approximately solving finite nuclei with realistic strong interactions. The NCSM and other approaches require an extrapolation of the results obtained in a finite basis space to the infinite basis space limit and assessment of the uncertainty of those extrapolations. Each observable requires a separate extrapolation and many observables have no proven extrapolation method. Herein, we propose a feed-forward artificial neural network (ANN) method as an extrapolation tool to obtain the ground-state energy and the ground-state point-proton root-mean-square (rms) radius along with their extrapolation uncertainties. The designed ANNs are sufficient to produce results for these two very different observables in 6 Li from the ab initio NCSM results in small basis spaces that satisfy the following theoretical physics condition: independence of basis space parameters in the limit of extremely large matrices. Comparisons of the ANN findings with other extrapolation methods are also provided.

6Li↗

DNN-based Signal Processing for Liquid Argon Time Projection Chambers

We investigate a deep learning-based signal processing for liquid argon time projection chambers (LArTPCs), a leading detector technology in neutrino physics. Identifying regions of interest (ROIs) in LArTPCs is challenging due to signal cancellation from bipolar responses and various detector effects observed in real data. We approach ROI identification as an image segmentation task, and employ a U-ResNet architecture. The network is trained on samples that incorporate detector geometry information and include a range of detector variations. Our approach significantly outperforms traditional methods while maintaining robustness across diverse detector conditions. This method has been adopted for signal processing in the Short-Baseline Neutrino program and provides a valuable foundation for future experiments such as the Deep Underground Neutrino Experiment.

Bhat, Avinay [Chicago U.]↗

Identifying Sample Provenance From SEM/EDS Automated Particle Analysis via Few-Shot Learning Coupled With Similarity Graph Clustering

Automated particle analysis (APA) provides a vast amount of compositional data via energy-dispersive X-ray spectroscopy along with size and shape data via scanning electron microscopy for individual particles in a sample. In many instances, APA data are leveraged to support identification of the source of a sample based on the detection of particles of a specific composition. Often, the particles that provide context make up a minuscule portion of the sample. Additionally, the interpretation of complex samples can be difficult due to the diversity of compositions both in the mixture and within a particle. In this work, we demonstrate a method to compute and cluster similarity graphs that describe inter-particle relationships within a sample using a multi-modal few-shot learning neural network. Here, as a proof-of-concept, we show that samples known to have been exposed to gunshot residue can be distinguished from samples occasionally mistaken for gunshot residue. Our workflow builds upon standard APA techniques and data processing methods to unveil additional information in a readily interpretable and quantitatively comparable format.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Real-Time Fault Tracking and Ground Motion Prediction for Large Earthquakes With HR-GNSS and Deep Learning

Earthquake early warning (EEW) systems aim to forecast the shaking intensity rapidly after an earthquake occurs and send warnings to affected areas before the onset of strong shaking. The system relies on rapid and accurate estimation of earthquake source parameters. However, it is known that source estimation for large ruptures in real-time is challenging, and it often leads to magnitude underestimation. In a previous study, we showed that machine learning, HR-GNSS, and realistic rupture synthetics can be used to reliably predict earthquake magnitude. This model, called Machine-Learning Assessed Rapid Geodetic Earthquake model (M-LARGE), can rapidly forecast large earthquake magnitudes with an accuracy of 99%. Here, we expand M-LARGE to predict centroid location and fault size, enabling the construction of the fault rupture extent for forecasting shaking intensity using existing ground motion models. We test our model in the Chilean Subduction Zone with thousands of simulated and five real large earthquakes. The result achieves an average warning time of 40.5 s for shaking intensity MMI4+, surpassing the 34 s obtained by a similar GNSS EEW model. Our approach addresses a critical gap in existing EEW systems for large earthquakes by demonstrating real-time fault tracking feasibility without saturation issues. This capability leads to timely and accurate ground motion forecasts and can support other methods, enhancing the overall effectiveness of EEW systems. Additionally, the ability to predict source parameters for real Chilean earthquakes implies that synthetic data, governed by our understanding of earthquake scaling, is consistent with the actual rupture processes.

58 GEOSCIENCES↗

Ensembles of Networks Produced from Neural Architecture Search

Neural architecture search (NAS) is a popular topic at the intersection of deep learning and high performance computing. NAS focuses on optimizing the architecture of neural networks along with their hyperparameters in order to produce networks with superior performance. Much of the focus has been on how to produce a single best network to solve a machine learning problem, but as NAS methods produce many networks that work very well, this affords the opportunity to ensemble these networks to produce an improved result. Additionally, the diversity of network structures produced by NAS drives a natural bias towards diversity of predictions produced by the individual networks. This results in an improved ensemble over simply creating an ensemble that contains duplicates of the best network architecture retrained to have unique weights.

Herron, Emily↗

Fast and accurate reduced-order modeling of a MOOSE-based additive manufacturing model with operator learning

One predominant challenge in additive manufacturing (AM) is to achieve specific material properties by manipulating manufacturing process parameters during the runtime. Such manipulation tends to increase the computational load imposed on existing simulation tools employed in AM. The goal of the present work is to construct a fast and accurate reduced-order model (ROM) for an AM model developed within the Multiphysics Object-Oriented Simulation Environment (MOOSE) framework, ultimately reducing the time/cost of AM control and optimization processes. Our adoption of the operator learning (OL) approach enabled us to learn a family of differential equations produced by altering process variables in the laser’s Gaussian point heat source. More specifically, we used the Fourier neural operator (FNO) and deep operator network (DeepONet) to develop ROMs for time-dependent responses. Furthermore, we benchmarked the performance of these OL methods against a conventional deep neural network (DNN)-based ROM. Ultimately, we found that OL methods offer comparable performance and, in terms of accuracy and generalizability, even outperform DNN at predicting scalar model responses. The DNN-based ROM afforded the fastest training time. Furthermore, all the ROMs were faster than the original MOOSE model yet still provided accurate predictions. FNO had a smaller mean prediction error than DeepONet, with a larger variance for time-dependent responses. Unlike DNN, both FNO and DeepONet were able to simulate time series data without the need for dimensionality reduction techniques. Finally, the present work can help facilitate the AM optimization process by enabling faster execution of simulation tools while still preserving evaluation accuracy.

36 MATERIALS SCIENCE↗

Neural network potential for Zr–Rh system by machine learning

Zr–Rh metallic glass has enabled its many applications in vehicle parts, sports equipment and so on due to its outstanding performance in mechanical property, but the knowledge of the microstructure determining the superb mechanical property remains yet insufficient. Here, we develop a deep neural network potential of Zr–Rh system by using machine learning, which breaks the dilemma between the accuracy and efficiency in molecular dynamics simulations, and greatly improves the simulation scale in both space and time. Here, the results show that the structural features obtained from the neural network method are in good agreement with the cases in ab initio molecular dynamics simulations. Furthermore, we build a large model of 5400 atoms to explore the influences of simulated size and cooling rate on the melt-quenching process of Zr 77 Rh 23 . Our study lays a foundation for exploring the complex structures in amorphous Zr 77 Rh 23 , which is of great significance for the design and practical application.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Classical analog of quantum models in synthetic dimensions

We introduce a classical analog of quantum matter in ultracold molecule-synthetic or Rydberg atom-synthetic dimensions, by extending the Potts model to include interactions J 1 between atoms adjacent in both real and synthetic space and studying its finite-temperature properties. For intermediate values of J 1 , the resulting phases and phase diagrams are similar to those of the clock and Villain models, in which three phases emerge. There exists a sheet phase analogous to that found in quantum synthetic dimension models between the high-temperature disordered phase and the low-temperature ferromagnetic phase. Furthermore, we also employ machine learning to uncover nontrivial features of the phase diagram using the learning by confusion approach, which is able to discern several successive phase transitions.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Adaptive Neuron Apoptosis for Accelerating Deep Learning on Large Scale Systems

Machine Learning and Data Mining (MLDM) algorithms are becoming ubiquitous in {\em model learning} from the large volume of data generated using simulations, experiments and handheld devices. Deep Learning algorithms -- a class of MLDM algorithms -- are applied for automatic feature extraction, and learning non-linear models for unsupervised and supervised algorithms. Naturally, several libraries which support large scale Deep Learning -- such as TensorFlow and Caffe -- have become popular. In this paper, we present novel techniques to accelerate the convergence of Deep Learning algorithms by conducting low overhead removal of redundant neurons -- {\em apoptosis} of neurons -- which do not contribute to model learning, during the training phase itself. We provide in-depth theoretical underpinnings of our heuristics (bounding accuracy loss and handling apoptosis of several neuron types), and present the methods to conduct adaptive neuron apoptosis. We implement our proposed heuristics with the recently introduced TensorFlow and using its recently proposed extension with MPI. Our performance evaluation on two difference clusters -- one connected with Intel Haswell multi-core systems, and other with nVIDIA GPUs -- using InfiniBand, indicates the efficacy of the proposed heuristics and implementations. Specifically, we are able to improve the training time for several datasets by 2-3x, while reducing the number of parameters by 30x (4-5x on average) on datasets such as ImageNet classification. For the Higgs Boson dataset, our implementation improves the accuracy (measured by Area Under Curve (AUC)) for classification from 0.88/1 to 0.94/1, while reducing the number of parameters by 3x in comparison to existing literature, while achieving a 2.44x speedup in comparison to the default (no apoptosis) algorithm.

Siegel, Charles M.↗

Adaptive Hierarchical Cyber Attack Detection and Localization in Active Distribution Systems

Development of a cyber security strategy for the active distribution systems is challenging due to the inclusion of distributed renewable energy generations. Here this paper proposes an adaptive hierarchical cyber attack detection and localization framework for distributed active distribution systems via analyzing electrical waveforms. Cyber attack detection is based on a sequential deep learning model, via which even minor cyber attacks can be identified. The two-stage cyber attack localization algorithm first estimates the cyber attack sub-region, and then localize the specified cyber attack within the estimated subregion. We propose a modified spectral clustering-based network partitioning method for the hierarchical cyber attack ‘coarse’ localization. Next, to further narrow down the cyber attack location, a normalized impact score based on waveform statistical metrics is proposed to obtain a ‘fine’ cyber attack location by characterizing different waveform properties. Finally, compared with classical and state-of-art methods, a comprehensive quantitative evaluation with two case studies shows promising estimation results of the proposed framework.

42 ENGINEERING↗

Accurate and transferable multitask prediction of chemical properties with an atoms-in-molecules neural network

Atomic and molecular properties could be evaluated from the fundamental Schrodinger’s equation and therefore represent different modalities of the same quantum phenomena. Here, we present AIMNet, a modular and chemically inspired deep neural network potential. We used AIMNet with multitarget training to learn multiple modalities of the state of the atom in a molecular system. The resulting model shows on several benchmark datasets state-of-the-art accuracy, comparable to the results of orders of magnitude more expensive DFT methods. It can simultaneously predict several atomic and molecular properties without an increase in the computational cost. With AIMNet, we show a new dimension of transferability: the ability to learn new targets using multimodal information from previous training. The model can learn implicit solvation energy (SMD method) using only a fraction of the original training data and an archive median absolute deviation error of 1.1 kcal/mol compared to experimental solvation free energies in the MNSol database.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Calorimetry with deep learning: particle simulation and reconstruction for collider physics

Using detailed simulations of calorimeter showers as training data, we investigate the use of deep learning algorithms for the simulation and reconstruction of single isolated particles produced in high-energy physics collisions. We train neural networks on single-particle shower data at the calorimeter-cell level, and show significant improvements for simulation and reconstruction when using these networks compared to methods which rely on currently-used state-of-the-art algorithms. We define two models: an end-to-end reconstruction network which performs simultaneous particle identification and energy regression of particles when given calorimeter shower data, and a generative network which can provide reasonable modeling of calorimeter showers for different particle types at specified angles and energies. We investigate the optimization of our models with hyperparameter scans. Furthermore, we demonstrate the applicability of the reconstruction model to shower inputs from other detector geometries, specifically ATLAS-like and CMS-like geometries. These networks can serve as fast and computationally light methods for particle shower simulation and reconstruction for current and future experiments at particle colliders.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗