检索结果-内蒙古大学图书馆

Machine Learning Analysis Using the Black Oil Model and Parallel Algorithms in Oil Recovery Forecasting

ALGORITHMS 2024年第8期17卷

作者： Matkerim, Bazargul Mukhanbet, Aksultan Kassymbek, Nurislam Daribayev, Beimbet Mustafin, Maksat Imankulov, Timur Natl Engn Acad Republ Kazakhstan Alma Ata 050010 Kazakhstan Al Farabi Kazakh Natl Univ Dept Comp Sci Alma Ata 050040 Kazakhstan Joldasbekov Inst Mech & Engn Alma Ata 050000 Kazakhstan

The accurate forecasting of oil recovery factors is crucial for the effective management and optimization of oil production processes. This study explores the application of machine learning methods, specifically focusing on parallel algorithms, to enhance traditional reservoir simulation frameworks using black oil models. This research involves four main steps: collecting a synthetic dataset, preprocessing it, modeling and predicting the oil recovery factors with various machine learning techniques, and evaluating the model's performance. The analysis was carried out on a synthetic dataset containing parameters such as porosity, pressure, and the viscosity of oil and gas. By utilizing parallel computing, particularly GPUs, this study demonstrates significant improvements in processing efficiency and prediction accuracy. While maintaining the value of the R2 metric in the range of 0.97, using data parallelism sped up the learning process by, at best, 10.54 times. neural network training was accelerated almost 8 times when running on a GPU. These findings underscore the potential of parallel machine learning algorithms to revolutionize the decision-making processes in reservoir management, offering faster and more precise predictive tools. This work not only contributes to computational sciences and reservoir engineering but also opens new avenues for the integration of advanced machine learning and parallel computing methods in optimizing oil recovery.

关键词： distributed machine learning HPC artificial intelligence cuML enhanced oil recovery

来源：评论

学校读者我要写书评

暂无评论

Human activity recognition in WBAN using ensemble model

引用

INTERNATIONAL JOURNAL OF PERVASIVE COMPUTING AND COMMUNICATIONS 2023年第4期19卷 513-549页

作者： Boga, Jayaram Kumar, Dhilip, V Vel Tech Rangarajan Dr Sagunthala R&D Inst Sci & Comp Sci & Engn Chennai Tamil Nadu India

Purpose For achieving the profitable human activity recognition (HAR) method, this paper solves the HAR problem under wireless body area network (WBAN) using a developed ensemble learning approach. The purpose of this study is,to solve the HAR problem under WBAN using a developed ensemble learning approach for achieving the profitable HAR method. There are three data sets used for this HAR in WBAN, namely, human activity recognition using smartphones, wireless sensor data mining and Kaggle. The proposed model undergoes four phases, namely, "pre-processing, feature extraction, feature selection and classification." Here, the data can be preprocessed by artifacts removal and median filtering techniques. Then, the features are extracted by techniques such as "t-distributed Stochastic Neighbor Embedding", "Short-time Fourier transform" and statistical approaches. The weighted optimal feature selection is considered as the next step for selecting the important features based on computing the data variance of each class. This new feature selection is achieved by the hybrid coyote Jaya optimization (HCJO). Finally, the meta-heuristic-based ensemble learning approach is used as a new recognition approach with three classifiers, namely, "support vector machine (SVM), deep neural network (DNN) and fuzzy classifiers." Experimental analysis is performed. Design/methodology/approach The proposed HCJO algorithm was developed for optimizing the membership function of fuzzy, iteration limit of SVM and hidden neuron count of DNN for getting superior classified outcomes and to enhance the performance of ensemble classification. Findings The accuracy for enhanced HAR model was pretty high in comparison to conventional models, i.e. higher than 6.66% to fuzzy, 4.34% to DNN, 4.34% to SVM, 7.86% to ensemble and 6.66% to Improved Sealion optimization algorithm-Attention Pyramid-Convolutional neural network-AP-CNN, respectively. Originality/value The suggested HAR model with WBAN using HCJO

关键词： Wireless body area network Human activity recognition Meta-heuristic-based ensemble learning Hybrid Coyote Jaya Optimization Deep neural network Fuzzy classifier Support vector machine

来源：评论

学校读者我要写书评

暂无评论

Edge Computing-Enabled Lightweight Deep neural network for Real-Time Video Surveillance in Maritime Cyber-Physical Systems 9

Edge Computing-Enabled Lightweight Deep Neural Network for R...

引用

9th International Conference on Intelligent Computing and Signal processing, ICSP 2024

作者： Luo, Caofei Wu, Fumin Xiong, Renhao Xu, Guanhua Liu, Wen Maritime Electronics Research Institute Co. Ltd. Ningbo315040 China College of Information Science and Electronic Engineering Zhejiang University Hangzhou310058 China School of Navigation Wuhan University of Technology Wuhan430063 China

ISBN: (纸本)9798350376548

With the rapid developments of low-end IoT devices and artificial intelligence (AI), the heterogeneous and dynamic data has been increasing drastically, especially in the era of AI-enabled maritime cyber-physical systems (CPS). It is intractable to provide the efficient and flexible computer vision applications through traditional cloud-centric data processing strategies, which commonly suffer from the limitations, e.g., high cost, high- latency communication, and high occupied storage, etc. To meet these challenges, edge computing has meaningfully evolved into a flexible and powerful distributed architecture. It contributes to improvements in bandwidth savings and response time in practical applications. To guarantee efficient video surveillance in AI-powered maritime CPS, the edge computing-based lightweight deep network (termed ShipNet-YOLOv3), incorporated with Kalman filtering and Hungarian algorithm, is proposed to implement real-time detection and tracking of moving ships. In particular, ShipNet-YOLOv3 adopts Darknet-53 as the backbone network (i.e., feature extractor) with three prediction boxes. The channel pruning and fine-tuning strategy are further introduced to balance the trade-off between computational cost, network size, and detection (or tracking) accuracy. Experimental results on both synthetic and realistic scenarios have illustrated that the proposed method could provide superior ship detection and tracking results under different adverse imaging conditions. It thus has the capacity of promoting real-time video surveillance in AI -enabled maritime CPS. © 2024 IEEE.

关键词： Kalman filters

来源：评论

学校读者我要写书评

暂无评论

Residual network Based Bidirectional Gated Recurrent Unit for Speech Recognition Using Speech Signals 2

Residual Network Based Bidirectional Gated Recurrent Unit fo...

引用

2nd International Conference on distributed Computing and Optimization Techniques, ICDCOT 2024

作者： Naga, Raju M. Jayachandran, T. Alzubaidi, Laith H. Mazumder, Debarshi Praveena, Hirald Dwaraka Gst Gitam Department of Cse Karnataka Nagadenahalli India Nandha Engineering College Department of Ece Tamilnadu Erode India The Islamic University Najaf Iraq Nitte Meenakshi Institute of Technology Department of Ai and Ds Bengaluru India School of Engineering Mohan Babu University Erstwhile Sree Vidya Nikethan Engineering College Department of Ece Andhra Pradesh India

ISBN: (纸本)9798350382952

Speech Recognition is challenging research ins speech processing field of research. Different method has available in previous research that attained high accuracy. However, accuracy of speech recognition is not increased because of irrelevant features in recognition stage. In this research, proposed a Residual network based Bidirectional Gated Recurrent Unit (ResNet based Bi-GRU) method for speech recognition using Libri Speech dataset. The signals from data are pre-processed and feature extraction is performed through Mel Frequency Cepstral Coefficients (MFCC) that extract relevant features. Then, the extracted features are fed to classification stage which recognized the speech signals with high accuracy and less error rates. Performance of developed method is evaluated to metrics of Word Error Rate (%), Character Error Rate (CER), accuracy and f1-score. Developed technique obtained less WER of 4.55%, CER of 4.1%, accuracy of 96.71% and f1-score of 96.15% that is effective than existing methods like Recurrent neural network (RNN) -Gated Recurrent Unit (GRU) and Long Short-Term Memory (LSTM). © 2024 IEEE.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

DAS-VSP NOISE ELIMINATION BASED ON THE DILATED PYRAMID ATTENTION network

引用

INTERNATIONAL JOURNAL OF INNOVATIVE COMPUTING INFORMATION AND CONTROL 2023年第5期19卷 1361-1375页

作者： Zhong, Tie Chen, Haoliang Dong, Shiqi Li, Jianpo Northeast Elect Power Univ Sch Elect Engn 169 Changchun Rd Jilin 132012 Peoples R China Northeast Elect Power Univ Sch Comp Sci 169 Changchun Rd Jilin 132012 Peoples R China Minist Water Resources Changjiang River Sci Res Inst Key Lab Geotech Mech & Engn 23 Huangpu Rd Wuhan 430010 Peoples R China

. distributed acoustic sensing (DAS) is widely used in seismic data acquisition due to its low cost and high acquisition density. However, the collected DAS records are often accompanied by various interferences, which can impact the recognition of effective information and subsequent processing. Traditional noise elimination methods have limitations and cannot fulfill the requirements. Therefore, deep learning-based methods have been introduced to solve the noise elimination problem of DAS records. However, these networks may have deficiencies in their structures, leading to a degradation in their denoising ability. To achieve denoising capabilities, we propose a multi-scale network named the dilated pyramid attention network (DPA-Net) to tackle complex DAS background noise. The DPA-Net uses pyramid-shaped modules to extract features at different scales, employs a serration dilated convolution in the pyramid module to suppress the checkerboard effect and increase the receptive field, and uses a multi-scale module to optimize the extraction of coarse features at low resolutions. Additionally, an attention module is employed to enhance the effective features, thus improving the denoising performance. Experiments on both synthetic and field records have verified that DPA-Net can effectively suppress complex DAS background noise and recover weak reflection signals.

关键词： Background noise suppression Convolutional neural network (CNN) Dis-tributed acoustic sensing (DAS) Vertical seismic profile (VSP)

来源：评论

学校读者我要写书评

暂无评论

Element-wise parallel deep learning for structural distributed damage diagnosis by leveraging physical properties of long-gauge static strain transmissibility under moving loads

引用

MECHANICAL SYSTEMS AND SIGNAL processing 2024年 220卷

作者： Liu, Yu-Song Yan, Wang-Ji Yuen, Ka-Veng Zhou, Wan-Huan Univ Macau State Key Lab Internet Things Smart City Taipa Peoples R China Univ Macau Dept Civil & Environm Engn Taipa Peoples R China Univ Macau Guangdong Hong Kong Macau Joint Lab Smart Cities Taipa Peoples R China

The Transmissibility Function (TF) has gained considerable interest in structural damage detection because of its relatively high sensitivity to damage and robustness to excitation. This study proposed a distributed damage diagnosis method for beam-like structures based on a long-gauge static strain TF defined as the ratio of the Fourier transform of static strain response under moving loads from a target long-gauge sensor to that from the reference sensor. It has been discovered that each long-gauge static strain TF is independent of moving loads, making it an ideal damage indicator. It exhibits direct proportionality to the ratio of bending stiffness between the two measured zones covered by the target and reference long-gauge strain sensors while being unaffected by the stiffness of any other zones not covered by these two sensors. Moreover, the TF at zero frequency was proved to be equivalent to the ratio of static strain time history areas, relying solely on the stiffness of the covered zones. Leveraged by the physical properties of long-gauge static strain TF, an element-wise parallel deep neural network architecture was developed to decompose damage diagnosis into independent sub-tasks on the element level, utilizing a variational autoencoder (VAE) in the Bayesian inference framework for extracting features from the long-gauge strain TF and a regressor for mapping the features to elemental stiffness reduction. The divide-and-conquer strategy and element-wise parallel learning architecture allow for a significant reduction in the number of labeled training samples generated based on the updated numerical baseline model as it only requires simulated scenarios involving a single damage element. The efficiency and robustness of the proposed method were demonstrated through a numerical simply supported beam and a laboratory experiment on a two-span bridge.

关键词： Transmissibility function Damage diagnosis Long-gauge strain sensing Physical property-guided method Moving load

来源：评论

学校读者我要写书评

暂无评论

Botnet Attacks Detection using Image processing Approach: Sequential neural network and Feature Selection 14

Botnet Attacks Detection using Image Processing Approach: Se...

引用

14th International Conference on Computing Communication and networking Technologies, ICCCNT 2023

作者： Putra, Taufiq Odhi Dwi Ahmad, Tohari Institut Teknologi Sepuluh Nopember Department of Informatics Surabaya Indonesia

ISBN: (纸本)9798350335095

Botnets are networks of compromised computers that are controlled by an attacker for malicious purposes such as spamming attacks, DDoS, or data theft. Botnet attacks detection is a challenging task due to their distributed and covert nature. Botnet attacks are a major security concern for organizations and individuals alike. In this paper we proposed a botnet attacks detection using image processing approach. We propose an AI model architecture using Sequential neural network and feature selection. There are two methods for converting raw data into image data. First method using predefined categorize for choosing features. The second method uses Pearson correlation value between features in dataset and the label. Predefined categorized features have successfully increased the performance of proposed AI model for botnet attacks detection as well as Pearson correlation value. © 2023 IEEE.

关键词： Botnet

来源：评论

学校读者我要写书评

暂无评论

Prototyping of Low-Cost Configurable Sparse neural processing Unit with Buffer and Mixed-Precision Reshapeable MAC Array 28

Prototyping of Low-Cost Configurable Sparse Neural Processin...

引用

IEEE 28th International Conference on Parallel and distributed Systems (IEEE ICPADS)

作者： Wu, Binyi Furtner, Wolfgang Waschneck, Bernd Mayr, Christian Tech Univ Dresden Fac Elect & Comp Engn Dresden Germany Infineon Technol Div Power & Sensor Syst Munich Germany Tech Univ Dresden Ctr Tactile Internet Human In The Loop CeTI Dresden Germany

ISBN: (纸本)9781665473156

More recently, it has become possible to run deep learning algorithms on edge devices such as microcontrollers due to continuous improvements in neural network optimization algorithms such as quantization and neural architecture search. Nonetheless, most of the embedded hardware available today still falls short of the requirements of running deep neural networks. As a result, specialized processors have emerged to improve the inference efficiency of deep learning algorithms. However, most are not for edge applications that require efficient and low-cost hardware. Therefore, we design and prototype a low-cost configurable sparse neural processing Unit (NPU). The NPU has a built-in buffer and a reshapable mixed-precision multiply-accumulator (MAC) array. The computing and memory resources of the NPU are parameterized, and different NPUs can be derived. Besides, users can also configure the NPU at runtime to fully utilize the resources. In our experiments, the 200MHz NPU with only 32 MACs is more than 32 times faster than the 400MHz STM32H7 when inferring MobileNet-V1. Besides, the yielded NPUs can achieve roofline or even beyond roofline performance. The buffer and reshapeable MAC array push the NPU's attainable performance to the roofline, while the feature of supporting sparsity allows the NPU to obtain performance beyond the roofline.

关键词： low-cost configurable neural processing unit mixed-precision sparsity

来源：评论

学校读者我要写书评

暂无评论

Peer-to-Peer Learning Dynamics of Wide neural networks

Peer-to-Peer Learning Dynamics of Wide Neural Networks

引用

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Shreyas Chaudhari Srinivasa Pranav Emile Anand José M. F. Moura Electrical and Computer Engineering Carnegie Mellon University School of Computer Science Georgia Institute of Technology

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Peer-to-peer learning is an increasingly popular framework that enables beyond-5G distributed edge devices to collaboratively train deep neural networks in a privacy-preserving manner without the aid of a central server. neural network training algorithms for emerging environments, e.g., smart cities, have many design considerations that are difficult to tune in deployment settings – such as neural network architectures and hyperparameters. This presents a critical need for characterizing the training dynamics of distributed optimization algorithms used to train highly nonconvex neural networks in peer-to-peer learning environments. In this work, we provide an explicit characterization of the learning dynamics of wide neural networks trained using popular distributed gradient descent (DGD) algorithms. Our results leverage both recent advancements in neural tangent kernel (NTK) theory and extensive previous work on distributed learning and consensus. We validate our analytical results by accurately predicting the parameter and error dynamics of wide neural networks trained for classification tasks.

关键词： Training Computer aided instruction Distance learning Heuristic algorithms neural networks Signal processing algorithms Prediction algorithms Peer-to-peer computing Classification algorithms Optimization

来源：评论

学校读者我要写书评

暂无评论

Energy efficient spiking neural network neuromorphic processing to enable decentralised service workflow composition in support of multi-domain operations 4

Energy efficient spiking neural network neuromorphic process...

引用

Conference on Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications IV

作者： Bent, Graham Simpkin, Christopher Li, Yuhua Preece, Alun Cardiff Univ Crime & Secur Res Inst Cardiff Wales

ISBN: (纸本)9781510651036;9781510651029

Future Multi-Domain Operations (MDO) will require the coordination of hundreds-even thousands-of devices and component services. This will demand the capability to rapidly discover the distributed devices/services and combine them into different workflow configurations, thereby creating the applications necessary to support changing mission needs. Motivated by neuromorphic processing models, in previous work it was shown that this can be achieved by using hyperdimensional symbolic semantic vector representations of the services/devices and workflows. Using a process of vector exchange the required services are dynamically discovered and inter-connected to achieve the required tasks. In network edge environments, the capability to perform these tasks with minimum energy consumption is critical. This paper describes how emerging spiking neural network (SNN) neuromorphic processing devices can be used to perform the required hyperdimensional vector computation (HDC) with significant energy savings compared to what can be achieved using traditional CMOS implementations.

关键词： AI Spiking neural networks Service Discovery Workflow Composition Hyperdimensional Computing Vector Symbolic Architecture

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：