Engineering Papers⌕ Search

Engineering topics

Wang, Dali

Publications and source records attributed to Wang, Dali.

At least 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

X-composer: enabling cross-environments in-situ workflows between HPC and cloud

As large-scale scientific simulations and big data analyses become more popular, it is increasingly more expensive to store huge amounts of raw simulation results to perform post-analysis. To minimize the expensive data I/O, "in-situ" analysis is a promising approach, where data analysis applications analyze the simulation generated data on the fly without storing it first. However, it is challenging to organize, transform, and transport data at scales between two semantically different ecosystems due to the distinct software and hardware difference. To tackle these challenges, we design and implement the X-Composer framework. X-Composer connects cross-ecosystem applications to form an "in-situ" scientific workflow, and provides a unified approach and recipe for supporting such hybrid in-situ workflows on distributed heterogeneous resources. X-Composer reorganizes simulation data as continuous data streams and feeds them seamlessly into the Cloud-based stream processing services to minimize I/O overheads. For evaluation, we use X-Composer to set up and execute a cross-ecosystem workflow, which consists of a parallel Computational Fluid Dynamics simulation running on HPC, and a distributed Dynamic Mode Decomposition analysis application running on Cloud. Our experimental results show that X-Composer can seamlessly couple HPC and Big Data jobs in their own native environments, achieve good scalability, and provide high-fidelity analytics for ongoing simulations in real-time.

Wang, Dali↗
DOE OSTI2021

Autonomous nondestructive evaluation of resistance spot welded joints

The application of non-destructive evaluation approaches has attracted strong interests in modern automotive industries. Here, we present an autonomous deep-computing framework to analyze raw videos from infrared systems and to predict weld nugget shape and size with unprecedented accuracy and speed. In a comprehensive training and testing experiment with 90 videos (seven sets of welding material stack-ups), a new method was developed to assemble sufficient datasets for neural network training. Our framework successfully predicts all the nugget shapes with F1 scores that range from 0.84 to 0.92. The total training time on Nvidia DGX station takes less than 10 min for each set of welding material stack-up. The real inference time of an individual dataset (with 30 video frames) takes about 0.005 s. The procedure and methods developed in the study can be applied to other image-based weld property prediction, as well as other manufacturing processes. Furthermore, our well-trained neural networks take limited memory resources (2.3 MB) and are suitable for embedded microprocessors for in-situ welding quality control as edge computing within an intelligent welding framework.

42 ENGINEERING↗
DOE OSTI2021

Development of Explainable, Knowledge-Guided AI Models to Enhance the E3SM Land Model Development and Uncertainty Quantification

Focal Area(s): (2)Predictive modeling using AI techniques and AI-derived model components; use of AI and other tools to design a prediction system comprising of a hierarchy of models. (3) Insight gleaned from complex data (both observed and simulated) using AI, big data analytics, and other advanced methods, including explainable AI and physics- or knowledge- guided AI. Science Challenge: The Energy Exascale Earth System Model (E3SM) is a fully coupled, state-of-the-science Earth system model that uses code optimized for DOE's advanced computers to address the most critical scientific questions facing our nation and society (Golaz et al., 2019). The E3SM Land model (ELM) is designed to understand how the changes in terrestrial land surfaces will interact with other Earth system components and has been used to understand hydrologic cycles, biogeophysics, and ecosystem dynamics. In spite of great successes, the ELM has several known issues that restrain rapid improvements. For example, the ELM uses equilibrium models to simulate dynamic land-climate interactions and it requires long model spin-up time to identify suitable initial conditions for transient simulations. The ELM lacks built-in uncertainty mechanisms that can improve the robustness of model predictions. The ELM is a holistic, deterministic model system with a rigid design, and in many situations, it is hard to modify the ELM system to incorporate new theory/hypothesis and new data across scales to address emerging science problems (such as predicting the impacts of water cycle extremes). In addition, The ELM is technically optimized for traditional CPU-centric computers and it cannot fully utilize the current and incoming leadership computers for model simulations and uncertainty quantification (UQ). The success of artificial intelligence (AI) has inspired scientists to use AI models to discover intrinsic features from simulation data (Chattopadhyay et al., 2020) and observational data (Reichstein et al., 2019) to gain further process understanding of Earth science problems. However, autonomous AI model training through deep learning usually requires a huge amount of annotated data. To overcome the limitations from the data and computing resources, knowledge-guided AI models are necessary where human-knowledge is ingested in model construction (Banino et al., 2018) and training process (Silver et al., 2016) for efficient learning. Herein, we present a new way that leverages the process understanding from the ELM to guide AI model development for the ELM enhancement and UQ. We hope this study can inspire further Earth and environmental system model developments and transformations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Robotic Understanding of Spatial Relationships Using Neural-Logic Learning

Understanding spatial relations of objects is critical in many robotic applications such as grasping, manipulation, and obstacle avoidance. Humans can simply reason object's spatial relations from a glimpse of a scene based on prior knowledge of spatial constraints. The proposed method enables a robot to comprehend spatial relationships among objects from RGB-D data. This paper proposed a neural-logic learning framework to learn and reason spatial relations from raw data by following logic rules on spatial constraints. The neural-logic network consists of three blocks: grounding block, spatial logic block, and inference block. The grounding block extracts high-level features from the raw sensory data. The spatial logic blocks can predicate fundamental spatial relations by training a neural network with spatial constraints. The inference block can infer complex spatial relations based on the predicated fundamental spatial relations. Simulations and robotic experiments evaluated the performance of the proposed method.

Wang, Dali↗
DOE OSTI2020

Resiliency-based restoration optimization for dependent network systems against cascading failures

Due to the increasing importance of large-scale and complex network systems and the potential for massive cascading failures in these real-world systems, modeling of system resiliency and optimization of restoration strategies to mitigate system performance loss caused by diverse disruptions is of significant interest among researchers and practitioners. Although society has experienced many incidents that demonstrate the influence of cascading failures aggravated by dependencies inside network systems, existing resiliency-based restoration optimization research rarely if ever jointly considers the impact of system dependencies on cascading failures. In this paper, different restoration prioritization strategies are applied to network systems subject to cascading failures that take into account system dependencies. By conducting case studies on synthetic networks and the U.S. airport network system, the effects of restoration strategies are evaluated using a system resiliency metric and two system performance measurements. Furthermore, the influence of system dependency characteristics and the interplay between them and restoration strategies on system resiliency regarding cascading failures are also investigated. This work demonstrates the distinct effects of restoration prioritization actions against cascading failures to mitigate the performance loss in network systems with different properties. It also provides insights about restoration improvement by considering dependency impacts to effectively reduce the intensity and extent of cascading failures.

42 ENGINEERING↗
DOE OSTI2020

Designing a parallel Feel-the-Way clustering algorithm on HPC systems

This paper introduces a new parallel clustering algorithm, named Feel-the-Way clustering algorithm, that provides better or equivalent convergence rate than the traditional clustering methods by optimizing the synchronization and communication costs. Our algorithm design centers on how to optimize three factors simultaneously: reduced synchronizations, improved convergence rate, and retained same or comparable optimization cost. To compare the optimization cost, we use the Sum of Square Error (SSE) cost as the metric, which is the sum of the square distance between each data point and its assigned clusters. Compared with the traditional MPI k-means algorithm, the new Feel-the-Way algorithm requires less communications among participating processes. As for the convergence rate, the new algorithm requires fewer number of iterations to converge. As for the optimization cost, it obtains the SSE costs that are close to the k-means algorithm. In the paper, we first design the full-step Feel-the-Way k-means clustering algorithm that can significantly reduce the number of iterations that are required by the original k-means clustering method. Next, we improve the performance of the full-step algorithm by adopting an optimized sampling-based approach, named reassignment-history-aware sampling. Our experimental results show that the optimized sampling-based Feel-the-Way method is significantly faster than the widely used k-means clustering method, and can provide comparable optimization costs. More extensive experiments with several synthetic datasets and real-world datasets (e.g., MNIST, CIFAR-10, ENRON, and PLACES-2) show that the new parallel algorithm can outperform the open source MPI k-means library by up to 110% on a high-performance computing system using 4,096 CPU cores. In addition, the new algorithm can take up to 51% fewer iterations to converge than the k-means clustering algorithm.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

High-Performance Computing for Earth System Modeling

High-performance computing (HPC) plays an important role during the development of Earth system models. This chapter reviews HPC efforts related to Earth system models, including community Earth system models and energy exascale Earth system models. Specifically, this chapter evaluates computational and software design issues, analyzes several current HPC-related model developments, and provides an outlook for some promising areas within Earth system modeling in the era of exascale computing.

Wang, Dali↗
DOE OSTI2020

Toward Large-Scale Image Segmentation on Summit

Semantic segmentation of images is an important computer vision task that emerges in a variety of application domains such as medical imaging, robotic vision and autonomous vehicles to name a few. While these domain-specific image analysis tasks involve relatively small image sizes (~ 10 2 × 10 2 ), there are many applications that need to train machine learning models on image data with extents that are orders of magnitude larger (~10 4 × 10 4 ). Training deep neural network (DNN) models on large extent images is extremely memory-intensive and often exceeds the memory limitations of a single graphical processing unit, a hardware accelerator of choice for computer vision workloads. Here, an efficient, sample parallel approach to train U-Net models on large extent image data sets is presented. Its advantages and limitations are analyzed and near-linear strong-scaling speedup demonstrated on 256 nodes (1536 GPUs) of the Summit supercomputer. Using a single node of the Summit supercomputer, an early evaluation of a recently released model parallel framework called GPipe is demonstrated to deliver ~ 2X speedup in executing a U-Net model with an order of magnitude larger number of trainable parameters than reported before. Performance bottlenecks for pipelined training of U-Net models are identified and mitigation strategies to improve the speedups are discussed. Together, these results open up the possibility of combining both approaches into a unified scalable pipelined and data parallel algorithm to efficiently train U-Net models with very large receptive fields on data sets of ultra-large extent images.

Seal, Sudip↗
DOE OSTI2020

OpenGraphGym: A Parallel Reinforcement Learning Framework for Graph Optimization Problems

This paper presents an open-source, parallel AI environment (named OpenGraphGym) to facilitate the application of reinforcement learning (RL) algorithms to address combinatorial graph optimization problems. This environment incorporates a basic deep reinforcement learning method, and several graph embeddings to capture graph features, it also allows users to rapidly plug in and test new RL algorithms and graph embeddings for graph optimization problems. This new open-source RL framework is targeted at achieving both high performance and high quality of the computed graph solutions. This RL framework forms the foundation of several ongoing research directions, including 1) benchmark works on different RL algorithms and embedding methods for classic graph problems; 2) advanced parallel strategies for extreme-scale graph computations, as well as 3) performance evaluation on real-world graph solutions.

Zheng, Weijian↗
DOE OSTI2019

DeepWelding: a Deep Learning Enhanced Approach to GTAW Using Multi-source Sensing Images

Deep learning has great potential to reshape manufacturing industries. In this paper, we present DeepWelding, a novel framework that applies deep learning techniques to improve gas tungsten arc welding (GTAW) process monitoring and penetration detection using multi-source sensing images. The framework is capable of analyzing multiple types of optical sensing images synchronously and consists of three deep learning enhanced consecutive phases: image preprocessing, image selection, and weld penetration classification. Specifically, we adopted generative adversarial networks (pix2pix) for image denoising and classic convolutional neural networks (AlexNet) for image selection. Both pix2pix and AlexNet delivered satisfactory performance. However, five individual neural networks with heterogeneous architectures demonstrated inconsistent generalization capabilities in the classification phase when holding out multi-source images generated with specific experimental settings. Therefore, two ensemble methods combining multiple neural networks were designed to improve the model performance on unseen data collected from different experimental settings. In conclusion, we also found that the quality of model prediction was heavily influenced by the data stream collection environment. We think these findings are beneficial for the broad intelligent welding community.

42 ENGINEERING↗
DOE OSTI2019

Towards Efficient Convolutional Neural Networks Through Low-Error Filter Saliency Estimation

Filter saliency based channel pruning is a state-of-the-art method for deep convolutional neural network compression and acceleration. This channel pruning method ranks the importance of individual filter by estimating its impact of each filter’s removal on the training loss, and then remove the least important filters and fine-tune the remnant network. In this work, we propose a systematic channel pruning method that significantly reduces the estimation error of filter saliency. Different from existing approaches, our method largely reduces the magnitude of parameters in a network by introducing alternating direction method of multipliers (ADMM) into the pre-training procedure. Therefore, the estimation of filter saliency based on Taylor expansion is significantly improved. Extensive experiments with various benchmark network architectures and datasets demonstrate that the proposed method has a much improved unimportant filter selection capability and outperform state-of-the-art channel pruning method.

Wang, Zi↗
DOE OSTI2019

XScan: An Integrated Tool for Understanding Open Source Community-Based Scientific Code

Many scientific communities have adopted community-based models that integrate multiple components to simulate whole system dynamics. The community software projects’ complexity, stems from the integration of multiple individual software components that were developed under different application requirements and various machine architectures, has become a challenge for effective software system understanding and continuous software development. The paper presents an integrated software toolkit called X-ray Software Scanner (in abbreviation, XScan) for a better understanding of large-scale community-based scientific codes. Our software tool provides support to quickly summarize the overall information of scientific codes, including the number of lines of code, programming languages, external library dependencies, as well as architecture-dependent parallel software features. The XScan toolkit also realizes a static software analysis component to collect detailed structural information and provides an interactive visualization and analysis of the functions. We use a large-scale community-based Earth System Model to demonstrate the workflow, functions and visualization of the toolkit. We also discuss the application of advanced graph analytics techniques to assist software modular design and component refactoring.

Zheng, Weijian↗
DOE OSTI2019

Parallel Computing for Module-Based Computational Experiment

Large-scale scientific code plays an important role in scientific researches. In order to facilitate module and element evaluation in scientific applications, we introduce a unit testing framework and describe the demand for module-based experiment customization. We then develop a parallel version of the unit testing framework to handle long-term simulations with a large amount of data. Specifically, we apply message passing based parallelization and I/O behavior optimization to improve the performance of the unit testing framework and use profiling result to guide the parallel process implementation. Finally, we present a case study on litter decomposition experiment using a standalone module from a large-scale Earth System Model. This case study is also a good demonstration on the scalability, portability, and high-efficiency of the framework.

Yao, Zhuo↗
DOE OSTI2019

Modeling Cell Migration with Convolutional Neural Network and Deep Reinforcement Learning

Cell migration modeling is a longstanding biological challenge, which is regulated by a highly complex set of regulatory mechanisms at multiple scales in a developmental system. This study presents a generic framework for regulatory mechanisms discovery during cell migration. This framework uses convolutional neural networks and reinforcement learning to better study navigation rules and mechanisms during cell migration. This framework adopts a flexible model-free approach that directly takes raw images as the sensory input. It can better handle simulation scenarios that involve cell division during embryogenesis. Computational experiments also prove that this model achieves better performance than a previous model with a fully connected neural network.

Wang, Zi↗
DOE OSTI2019

Augmenting C. elegans Microscopic Dataset for Accelerated Pattern Recognition

The detection of cell shape changes in 3D time-lapse images of complex tissues is an important task. However, it is a challenging and tedious task to establish a comprehensive dataset to improve the performance of deep learning models. In the paper, we present a deep learning approach to augment 3D live images of the Caenorhabditis elegans embryo, so that we can further speed up the specific structure pattern recognition. We use an unsupervised training over unlabeled images to generate supplementary datasets for further pattern recognition. Technically, we used Alex-style neural networks in a generative adversarial network framework to generate new datasets that have common features of the C. elegans membrane structure. We also made the dataset available for a broad scientific community.

Lu, Zheng↗
DOE OSTI2018

Optimization of stepwise clustering algorithm in backward trajectory analysis

In recent years, the backward trajectory model has been widely used in the research of meteorological and atmospheric environmental quality. This paper presents a comprehensive study on a stepwise clustering analysis algorithm in the clustering process of backward trajectory model and an application of the clustering analysis of single-particle backward trajectory in 2016 in Changchun City. This study starts with an analysis of the original stepwise clustering algorithm and its application to a clustering process of 8784 backward trajectories during 48 h in Changchun City as a benchmark test case. Then, two improvements are made in the algorithm: First, in the process of finding the optimal classification, the algorithm complexity is improved from original O(n 3 ) to O(log(n)*n 2 ) through algorithm improvement. The algorithm performance is enhanced by log(n) times. Next, in the process of re-establishing the classification, the algorithm complexity is improved from the original O(m*n 2 ) to O(m*log(n)*n), that is another algorithm performance improvement by a factor of log(n). Therefore, the accumulative execution efficiency improvement through the algorithm optimization is 2*log(n) times, which has been further verified in the practical application in Changchun City.

97 MATHEMATICS AND COMPUTING↗