检索结果-内蒙古大学图书馆

Carbon dioxide storage and cumulative oil production predictions in unconventional reservoirs applying optimized machine-learning models

引用

Petroleum science 2025年第1期22卷 296-323页

作者： Shadfar Davoodi Hung Vo Thanh David A.Wood Mohammad Mehrad Sergey V.Muravyov Valeriy S.Rukavishnikov School of Earth Sciences & Engineering Tomsk Polytechnic University Laboratory for Computational Mechanics Institute for Computational Science and Artificial Intelligence Van Lang University Applied Science Research Center Applied Science Private University DWA Energy Limited Division for Automation & Robotics Tomsk Polytechnic University

To achieve carbon dioxide（CO2） storage through enhanced oil recovery, accurate forecasting of CO2subsurface storage and cumulative oil production is essential. This study develops hybrid predictive models for the determination of CO2storage mass and cumulative oil production in unconventional reservoirs. It does so with two multi-layer perceptron neural networks（MLPNN） and a least-squares support vector machine（LSSVM）, hybridized with grey wolf optimization（GWO） and/or particle swarm optimization（PSO）. Large, simulated datasets were divided into training（70%） and testing（30%）groups, with normalization applied to both groups. Mahalanobis distance identifies/eliminates outliers in the training subset only. A non-dominated sorting genetic algorithm（NSGA-II） combined with LSSVM selected seven influential features from the nine available input parameters: reservoir depth, porosity,permeability, thickness, bottom-hole pressure, area, CO2injection rate, residual oil saturation to gas flooding, and residual oil saturation to water flooding. Predictive models were developed and tested,with performance evaluated with an overfitting index（OFI）, scoring analysis, and partial dependence plots（PDP）, during training and independent testing to enhance model focus and effectiveness. The LSSVM-GWO model generated the lowest root mean square error（RMSE） values（0.4052 MMT for CO2storage and 9.7392 MMbbl for cumulative oil production） in the training group. That trained model also exhibited excellent generalization and minimal overfitting when applied to the testing group（RMSE of0.6224 MMT for CO2storage and 12.5143 MMbbl for cumulative oil production）. PDP analysis revealed that the input features “area” and “porosity” had the most influence on the LSSVM-GWO model's prediction performance. This paper presents a new hybrid modeling approach that achieves accurate forecasting of CO2subsurface storage and cumulative oil production. It also establishes a new standard for such forecasting,

关键词： Hybrid machine learning Least-squares support vector machine Grey wolf optimization Feature selection Carbon dioxide storage Enhanced oil recovery

来源：评论

学校读者我要写书评

暂无评论

Effect of order–disorder transition on thermodynamic and electronic properties of σ and χ phases in W–Re alloy: the first-principles calculation

引用

Tungsten 2025年第1期7卷 195-213页

作者： Zhi-Peng Zhu Dian Jia William Yi Wang Jun-Lei Yin Xing-Yu Gao Shu-Feng Yang Hai-Feng Song Jin-Shan Li State Key Laboratory of Solidification Processing Northwestern Polytechnical University Laboratory of Computational Physics Institute of Applied Physics and Computational Mathematics School of Metallurgical and Ecological Engineering University of Science and Technology Beijing

The Tungsten-Rhenium(W-Re) alloys,celebrated for their high melting point,strength at elevated temperatures,electrical resistivity,and radiation resistance,have been widely utilized in high-temperature components,aerospace,electronics,and nuclear *** constituents of the topologically close-packed(TCP) phases,the sigma phase(σ) and chi phase(χ) formed within W-Re alloys wield considerable influence on the mechanical properties and the stability of the *** calculations were utilized in the present work to explore the structural,thermodynamic,and electronic properties of both ordered and disordered configurations within the σ and χ phases,culminating in a systematic elucidation of the higher phase stability exhibited by the ordered *** is found that the bulk modulus of these two phases is directly proportional to the concentration of Re in the alloy,while the equilibrium volume is inversely *** thermodynamic parameters of the σ and χ phases are calculated via the mean-field potential *** similar trends observed in the isobaric heat capacity,enthalpy increment,and entropy change curves for these two phases suggest they possess comparable thermodynamic *** is noteworthy that the contribution of ionic vibrations predominantly affects the isobaric heat capacity,while the contribution of thermal electronic excitations increases linearly with *** the structure and thermodynamic properties of TCP phases in W-Re alloys at low temperatures has profound significance for optimizing material performance,microstructures features,establishing theoretical foundations,and predicting material behavior.

关键词： First-principles calculations Phase stability Thermodynamic properties W-Re alloy TCP phases

来源：评论

学校读者我要写书评

暂无评论

Ensemble framework for concept drift detection and class imbalance in data streams

引用

Multimedia Tools and Applications 2025年第11期84卷 8823-8837页

作者： Priya, S. Annie Uthra, R. Department of Computational Intelligence College of Engineering and Technology SRM Institute of Science and Technology Tamil Nadu Kattankulathur603 203 India

Many data mining application generate data in the form of streams called as streaming data and they arrive continuously. The distribution of data changes over time for the streaming data. The online ensemble learning method is used to handle the change in the underlying distribution of data called as concept drift, which gives the timely response for incoming data instances. Although many methods have been proposed for concept drift detection, data streams pose challenge in learning from concept drift with class imbalance, which exist in real world application such as intrusion detection and fault detection. Therefore, it is a significant challenge for the machine learning community to learn from the drifting and imbalanced data stream. In this paper, Ensemble Based Enhanced Early Drift Detection Model and Random Resampling Technique for online learning is proposed which detects the drift based on average error rate and standard deviation. The class imbalance is handled in the data stream by generating the synthetic data using the random resampling technique and the concept drift adaptation is done using ensemble classifiers. The proposed ensemble method can handle both concept drift and class imbalance with 98.52% accuracy. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

Surge-protected mechanical amplifier of ion transport devised using a bipolar nanopore with broken symmetry

引用

Physical Review E 2025年第2期111卷 L023101-L023101页

作者： Arghyadeep Paul N. R. Aluru Department of Mechanical Science and Engineering Beckman Institute for Advanced Science and Technology University of Illinois at Urbana−Champaign Urbana Illinois 61801 United States Walker Department of Mechanical Engineering Oden Institute for Computational Engineering and Sciences The University of Texas at Austin Austin Texas 78712 United States

Electrokinetic ion transport under coupled external electric field and pressure drops has recently exhibited giant, but monotonous, mechanical amplifications of ionic current in some nanochannels. While such effects can be exploited to develop novel pressure-sensitive ionic circuits, they can also yield large undesirable current surges if the control parameter, i.e., the pressure drop, overshoots. To overcome this critical issue, we design a nanopore that displays a similar mechanical activation of ionic current, but up to a limited magnitude of pressure, after which current ceases to grow. We integrate a bipolar surface charge distribution with an asymmetric channel geometry in a nanopore to alter the effect of an applied pressure drop on the quasistatic distribution of ions in the pore, which directs how the ionic conductance varies with pressure, in a way that yields the desired current-pressure response. Via such steps, analyzed using numerical simulations, we devise a funnel-shaped nanopore with a negatively charged conical mouth and a positively charged elongated cylindrical stem, that acts as the intended “surge-protected” mechanical amplifier of ion transport. It yields a substantial pressure-driven growth in ionic current (by factors of ∼4) up to a limited intensity of pressure drop exerted (along a constant external electric field) toward the stem, beyond which current begins to fall. Besides exhibiting this hitherto undetected electrokinetic behavior, our study implements and paves the way to an innovative “function to design” approach in the field of electrokinetics that can help build improved nanofluidic devices with desired advanced ion transport functionalities.

关键词： Nanopores

来源：评论

学校读者我要写书评

暂无评论

Hybrid RMDL-CNN for speech recognition from unclear speech signal

引用

International Journal of Speech Technology 2025年第1期28卷 195-217页

作者： Bhargava, Raja Arivazhagan, N. Babu, Kunchala Suresh Research Scholar Department of Computer Science and Engineering SRM Institute of Science and Technology Kattankulathur Chennai India Department of Computational Intelligence SRM Institute of Science and Technology Kattankulathur Chennai India Department of Computer Science and Engineering Potti Sriramulu Chaluvadi Mallikarjuna Rao college of Engineering and Technology Andhra Pradesh Vijayawada India

ASR is an effectual approach, which converts human speech into computer actions or text format. It involves extracting and determining the noise feature, the audio model, and the language model. The extraction and determination of the noise feature is a crucial aspect of speech recognition, serving as both a process of information compression and signal deconvolution. ASR schemes are mostly employed in smart homes, smart appliances, and biometric schemes. Yet, traditional approaches offer very low performance because of a noisy environment. Moreover, local differences and accents negatively influence the ASR scheme execution during the conversion of the speech signals. This paper introduces a hybrid RMDL-CNN method to address these challenges. At first, the input of unclear speech is carried out by the dataset. Then, signal pre-processing is done by employing a Gaussian filter. After that, voice enhancement is accomplished by employing nonlinear spectral subtraction. Later, the speech word is segmented from the enhanced output based on the Attentional Encoder-Decoder approach and finally, the speech is recognized using the proposed RMDL-CNN. The RMDL-CNN method is devised by the combination of RMDL and CNN. Furthermore, the established RMDL-CNN is accessed for its efficiency based on several values of k-group value, as well as learning data. In addition, the introduced RMDL-CNN approach for speech recognition achieved better accuracy, PPV, as well as NPV of 0.909, 0.947, and 0.917 for dataset 1. Moreover, the RMDL-CNN has achieved the highest accuracy of 0.909, PPV of 0.926 and NPV of 0.888 for dataset 2. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Cervical Cancer Risk Prediction using SMOTE and Explainable AI

Cervical Cancer Risk Prediction using SMOTE and Explainable ...

引用

2025 International Conference on Pervasive computational Technologies, ICPCT 2025

作者： Agarwal, Priyansh Chaudhary, Annanya Sarvanan, T.R. Srm Institute of Science and Technology School of Computing Department of Computational Intelligence India Srm Institute of Science and Technology Faculty of Engineering & Tech. Department of Computational Intelligence India

ISBN: (纸本)9798331508685

Cervical cancer, accounting for 7.9% of cancers in women globally, is a critical health challenge, particularly due to its asymptomatic nature in early stages. This study proposes a machine learning-driven framework to predict cervical cancer risk using clinical data, addressing class imbalance through SMOTE and optimizing model performance via GridSearchCV. By evaluating five classifiers - K-Nearest Neighbors, Random Forest, Support Vector Classifier, Logistic Regression, and Decision Trees - this research identifies the Decision Tree model as the most robust, achieving superior precision, recall, and accuracy metrics. Additionally, Explainable AI techniques such as SHAP and LIME were employed to enhance interpretability, offering clinicians actionable insights into key risk factors driving predictions. The results demonstrate the effectiveness of machine learning in facilitating early detection and personalized treatment strategies for cervical cancer, thereby advancing its integration into clinical applications. © 2025 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Fractal autoencoder with redundancy regularization for unsupervised feature selection

引用

science China(Information sciences) 2025年第2期68卷 89-102页

作者： Meiting SUN Fangyu LI Honggui HAN School of Information Science and Technology Beijing University of Technology Beijing Key Laboratory of Computational Intelligence and Intelligent System Engineering Research Center of Digital Community Ministry of Education Beijing Institute of Artificial Intelligence

Feature selection is a crucial step in data preprocessing because feature selection reduces the dimensionality of data by eliminating irrelevant and redundant features. Since manual labeling is expensive, unsupervised feature selection has received increasing attention in recent years. However, existing unsupervised feature selection methods tend to prioritize selecting highly correlated features over exploring feature diversity. Thus, a regularized fractal autoencoder(RFAE) method is proposed to select informative features in an unsupervised way. Specifically, the fractal autoencoder network extends autoencoders to construct a correspondence neural network and a selection neural network. The correspondence neural network exploits interfeature correlations and the selection neural network selects the informative features. A redundancy regularization strategy consists of a redundancy elimination regularization term based on the dependency between features and a sparse regularization term based on the group lasso. The redundancy regularization strategy eliminates feature subset redundancy and enhances network generalization ability. Extensive experimental results on six publicly available datasets show that the proposed RFAE outperforms the compared methods regarding clustering accuracy and classification accuracy. Moreover, the proposed RFAE achieves acceptable computation efficiency.

关键词： unsupervised feature selection fractal autoencoder correspondence neural network selection neural network redundancy regularization strategy

来源：评论

学校读者我要写书评

暂无评论

Chemical reaction mechanisms and models of energetic materials: A perspective

Energetic Materials Frontiers

引用

Energetic Materials Frontiers 2025年第1期6卷 129-144页

作者： Meng, Li Song, Qing-guan Yao, Chuang Zhang, Lei Pang, Si-ping School of Materials Science & Engineering Beijing Institute of Technology Beijing100081 China Institute of Applied Physics and Computational Mathematics Beijing100088 China Institute of Chemical Materials China Academy of Engineering Physics Mianyang621999 China of Chongqing School of Materials Science and Engineering Yangtze Normal University Chongqing408100 China

Energetic materials (EMs) are a kind of metastable functional materials with certain potential barriers, overcoming which can quickly release the energy stored in EMs. A thorough understanding of reaction mechanisms and accurate quantification of reaction rates are fundamental issues for optimizing energy output, ensuring hazard mitigation, and assessing the safety levels of EMs. This perspective provides an overview of research progress in chemical reaction mechanisms and models, with a particular emphasis on organic EMs and reactive metals. Organic EMs are mainly composed of carbon, hydrogen, nitrogen, and oxygen elements, enabling supersonic and self-sustaining detonation reactions capable of significant energy output. The incorporation of reactive metals like aluminum, magnesium, and boron has been recently found to augment the combustion heat and explosion temperature of EM formulations, sparking heightened research interest. This perspective first presents both EMs’ reaction mechanisms revealed via multiscale simulations and experimental methods, including thermal decomposition, shock initiation, and post combustion. Then, quantitatively characterized expressions of the physical models derived from the revealed mechanisms, including mathematical expressions like elementary and phenomenological reaction kinetic models, and emerging data-driven machine learning models, are reviewed. Finally, the view of the application, existing problems, and further development directions are outlined. © 2024 The Authors

关键词： Reaction rates

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Survey on Enhancing Digital Twin Security Systems with Explainable AI Techniques 3

A Comprehensive Survey on Enhancing Digital Twin Security Sy...

引用

3rd International Conference on Intelligent Data Communication Technologies and Internet of Things, IDCIoT 2025

作者： Lalithadevi, B. Krishnaveni, S. Department of Computer Science and Engineering Sathyabama Institute of Science and Technology Chennai India Department of Computational Intelligence SRM Institute of Science and Technology Kattankulathur India

ISBN: (纸本)9798331527549

Digital Twin (DT) technology has emerged as a cornerstone in modern engineering, enabling real-time digital replication of physical assets for monitoring, diagnostics, and optimization. However, as DTs become increasingly integrated into critical infrastructure, their security has become a paramount concern. This paper presents a comprehensive survey on the integration of Explainable Artificial Intelligence (XAI) into Digital Twin security, highlighting the potential of XAI in enhancing the transparency, interpretability, and trustworthiness of AI - driven security mechanisms in DTs. We explore various XAI techniques, including model- agnostic approaches, interpretable neural networks, and rule-based systems, detailing their potential applications in bolstering the security of Digital Twins. Furthermore, we identify several key challenges in implementing XAI for DT security, such as the need for standardized metrics for interpretability, the trade-offs between model complexity and explainability, and the potential for adversarial attacks on explainable models. Through our analysis, we highlight the critical research gaps, including the security of Digital Twin (DT) systems and explores the integration of Explainable AI (XAI) to enhance transparency, interpretability, and trustworthiness. Outcomes include detailed XAI applications for DT security, identification of limitations, and a roadmap for future research. We discuss various XAI techniques, their applications, and the current challenges in implementing XAI for DT security. Our survey identifies key research gaps and offers a roadmap for future research in this emerging field. © 2025 IEEE.

关键词： Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Learning Maximally Spanning Representations Improves Protein Function Annotation 29th

Learning Maximally Spanning Representations Improves Protein...

引用

29th International Conference on Research in computational Molecular Biology, RECOMB 2025

作者： Luo, Jiaqi Luo, Yunan School of Computational Science and Engineering Georgia Institute of Technology AtlantaGA United States

ISBN: (纸本)9783031902512

Protein function prediction is an essential problem in the field of computational biology and has a significant influence on many applications like drug discovery and antibody design. A long-lasting challenge for this problem is the data imbalance issue in existing protein function data, known as the long-tail problem: a small set of function classes are associated with most of the annotated proteins, while the majority of classes only have very few labeled proteins. This leads to the biased performance of current machine learning-based protein function prediction models, which focus more on optimizing prediction accuracy for well-represented function classes while performing poorly for underrepresented classes. In this work, we developed MSRep, a deep learning-based framework for protein function annotation aimed to mitigate the data imbalance issue. The inspiration for MSRep comes from an interesting phenomenon, called neural collapse (NC), observed in high-accuracy deep classification models. MSRep leveraged an NC-inspired loss to refine a pretrained protein language model embedding into a new space with NC properties, where both the well- and under-represented classes are equally represented in the embedding space. Evaluations across four protein function annotation tasks suggested that MSRep provides accurate predictions for both well- and underrepresented classes, outperforming several state-of-the-art annotation tools. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Deep learning

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：