检索结果-南通市图书馆

A benchmark-based method for evaluating hyperparameter optimization techniques of neural networks for surface water quality prediction

引用

Frontiers of Environmental Science & Engineering 2024年第5期18卷 13-27页

作者： Xuan Wang Yan dong Jing Yang Zhipeng liu Jinsuo Lu School of Environmental and Municipal Engineering Xi’an University of Architecture and TechnologyXi’an710055China Shaanxi Key Laboratory of Environmental Engineering Xi’an University of Architecture and TechnologyXi’an710055China State Key Laboratory of Green Building in West China Xi’an University of Architecture and TechnologyXi’an710055China

Neural networks(NNs)have been used extensively in surface water prediction tasks due to computing algorithm improvements and data *** essential step in developing an NN is the hyperparameter *** practice,it is common to manually determine hyperparameters in the studies of NNs in water resources *** may result in considerable randomness and require significant computation time;therefore,hyperparameter optimization(HPO)is *** study adopted five representatives of the HPO techniques in the surface water quality prediction tasks,including the grid sampling(GS),random search(RS),genetic algorithm(GA),Bayesian optimization(BO)based on the Gaussian process(GP),and the tree Parzen estimator(TPE).For the evaluation of these techniques,this study proposed a method:first,the optimal hyperparameter value sets achieved by GS were regarded as the benchmark;then,the other HPO techniques were evaluated and compared with the benchmark in convergence,optimization orientation,and consistency of the optimized *** results indicated that the TPE-based BO algorithm was recommended because it yielded stable convergence,reasonable optimization orientation,and the highest consistency rates with the benchmark *** optimization consistency rates via TPE for the hyperparameters hidden layers,hidden dimension,learning rate,and batch size were 86.7%,73.3%,73.3%,and 80.0%,*** the evaluation of HPO techniques directly based on the prediction performance of the optimized NN in a single HPO test,the proposed benchmark-based HPO evaluation approach is feasible and robust.

关键词： Neural networks Hyperparameter optimization Surface water quality prediction Bayes optimization Genetic algorithm

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Data and Model Driven Task Offloading Strategy in the Dynamic Mobile Edge Computing System

引用

Journal of Systems Science & Complexity 2024年第1期37卷 351-368页

作者： dong Hairong WU Wei SONG Haifeng liu Zhen ZHANG Zixuan School of Electronic and Information Engineering Beijing Jiaotong UniversityBeijing 100044China School of Electronic and Information Engineering BeiHang UniversityBeijing 100191China CRSC Research&Design Institute Group Co. Ltd.Beijing 100070China

Mobile Edge Computing(MEC)provides communication and computational capabilities for the industrial Internet,meeting the demands of latency-sensitive ***,traditional model-driven task offloading strategies face challenges in adapting to situations with unknown network communication status and computational *** limitation becomes notably significant in complex industrial networks of high-speed *** by these considerations,a data and model-driven task offloading problem is proposed in this paper.A redundant communication network is designed to adapt to anomalous channel states when tasks are offloaded to edge *** link switching mechanism is executed by the train according to the attributes of the completed *** task offloading optimization problem is formulated by introducing data-driven prediction of communication states into the traditional ***,the optimal strategy is achieved by employing the informer-based prediction algorithm and the quantum particle swarm optimization method,which effectively tackle real-time optimization problems due to their low time *** simulations illustrate that the data and model-driven task offloading strategy can predict the communication state in advance,thus reducing the cost of the system and improving its robustness.

关键词： Data driven model informer mobile edge computing quantum particle swarm optimization task offloading.

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Discovering Latent Variables for the Tasks With Confounders in Multi-Agent Reinforcement Learning

引用

IEEE/CAA Journal of Automatica Sinica 2024年第7期11卷 1591-1604页

作者： Kun Jiang wenzhang liu Yuanda Wang Lu dong Changyin Sun the School of Automation Southeast UniversityNanjing 210096China the School of Artificial Intelligence Anhui UniversityHefei 230601China IEEE the School of Cyber Science and Engineering Southeast UniversityNanjing 211189China the Engineering Research Center of Autonomous Unmanned System Technology Ministry of EducationHefei 230601China

Efficient exploration in complex coordination tasks has been considered a challenging problem in multi-agent reinforcement learning(MARL). It is significantly more difficult for those tasks with latent variables that agents cannot directly observe. However, most of the existing latent variable discovery methods lack a clear representation of latent variables and an effective evaluation of the influence of latent variables on the agent. In this paper, we propose a new MARL algorithm based on the soft actor-critic method for complex continuous control tasks with confounders. It is called the multi-agent soft actor-critic with latent variable(MASAC-LV) algorithm, which uses variational inference theory to infer the compact latent variables representation space from a large amount of offline ***, we derive the counterfactual policy whose input has no latent variables and quantify the difference between the actual policy and the counterfactual policy via a distance function. This quantified difference is considered an intrinsic motivation that gives additional rewards based on how much the latent variable affects each agent. The proposed algorithm is evaluated on two collaboration tasks with confounders, and the experimental results demonstrate the effectiveness of MASAC-LV compared to other baseline algorithms.

关键词： Latent variable model maximum entropy multi-agent reinforcement learning(MARL) multi-agent system

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Directional amine-based solvent extraction for simultaneous enhanced water recovery,salt separation and effective descaling from hypersaline brines

引用

Frontiers of Environmental Science & Engineering 2024年第3期18卷 83-92页

作者： Yuyao Zhang Litao Jia Jin Zhao Xuming liu Shuyu dong Chuanyang liu Yuanyuan Cui Shougang Research Institute of Technology Beijing 100043China Research Center of Environmental Technology in Water Pollution Source Identification and Precise Supervision School of EnvironmentTsinghua UniversityBeijing 100084China

The feasibility of simultaneous water recovery,salt separation and effective descaling of hypersaline brine was investigated by diisopropylamine(DIPA)-based directional solvent extraction(DSE),using diluted/concentrated seawater with initial saline concentration range of 12–237 g/L at extraction temperatures of 5 and 15℃,*** water recovery shows an obvious boundary at saline concentration of 115 g/L under dual effect of specific water extraction efficiency and extraction *** Cl^(–) ion concentration in product water is in sharp contrast to the nearly complete removal of SO_(4)^(2–)and hardness ions,indicating that DIPA-based DSE process indeed achieved efficient separation and purification of Cl^(–) ion from hypersaline ***,the radical precipitation of Mg^(2+)and Ca^(2+)ions in form of Mg(OH)_(2)and CaCO_(3)demonstrates effective descaling potential,although it leads to more DIPA residues in dewatered raffinate than product ***,an exponential correlation between the Cl^(–) removal efficiency and specific water extraction efficiency further reveals the intrinsic relationship of water extraction process and transfer of Cl^(–) ion to the product ***,the study provides a novel approach for integrating the water recovery and separation of Cl^(–) ion from ultra-high-salinity brines with radical precipitation of Mg^(2+)and Ca^(2+) ions in one step.

关键词： Hypersaline brines Directional solvent extraction Salt separation Ca^(2+)and Mg^(2+)descaling

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Comparison of Direct Pore-Scale and Volume-Averaging Methods for the Performance Evaluation of Porous Volumetric Solar Receiver

引用

Journal of Thermal Science 2024年第5期33卷 1607-1617页

作者： DU Shen HE Yaling LI dong liu Zhanbin LI Mengjie National Innovation Platform(Center)for Industry-Education Integration of Energy Storage Technology Key Laboratory of Thermo-Fluid Science and Engineering of Ministry of EducationXi’an Jiaotong UniversityXi’an710049China

Direct pore-scale and volume-averaging numerical simulations are two methods for investigating the performance of porous volumetric solar *** clarify the difference in the prediction of heat transfer processes,a direct comparison between these two methods was conducted at both steady state and transient *** numerical models were established based on X-ray computed tomography scans and a local thermal non-equilibrium model,*** empirical parameters,which are indispensable to the volume-averaging simulation,were determined by Monte Carlo ray tracing and direct pore-scale numerical *** predicted outlet air temperature of the receiver by the volume-averaging simulation method corresponded satisfactorily to that in the direct pore-scale *** largest discrepancies were observed when the receiver's working temperature was elevated,with differences of 5.5%and 3.68%for the steady state and transient state simulations,***,the volume-averaging method is incapable of capturing the local temperature information of the air and porous *** underestimates the inlet temperature of the receiver,leading to an overestimation of the receiver's thermal efficiency,with the largest difference being 6.51%.The comparison results show that the volume-averaging model is a good approximation to the pore-scale model when the empirical parameters are carefully selected.

关键词： porous media volumetric solar receiver pore-scale method volume-averaging method thermal performance

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Learning-based counterfactual explanations for recommendation

引用

Science China(Information Sciences) 2024年第8期67卷 115-129页

作者： Jingxuan wen Huafeng liu Liping JING Jian YU School of Computer and Information Technology Beijing Jiaotong University Beijing Key Lab of Traffic Data Analysis and Mining Beijing Jiaotong University

Counterfactual explanations provide explanations by exploring the changes in effect caused by changes in cause. They have attracted significant attention in recommender system research to explore the impact of changes in certain properties on the recommendation mechanism. Among several counterfactual recommendation methods, item-based counterfactual explanation methods have attracted considerable attention because of their flexibility. The core idea of item-based counterfactual explanation methods is to find a minimal subset of interacted items(i.e., short length) such that the recommended item would topple out of the top-K recommendation list once these items have been removed from user interactions(i.e., good quality). Usually, explanations are generated by ranking the precomputed importance of items, which fails to characterize the true importance of interacted items due to separation from the explanation generation. Additionally, the final explanations are generated according to a certain search strategy given the precomputed importance. This indicates that the quality and length of counterfactual explanations are deterministic;therefore, they cannot be balanced once the search strategy is fixed. To overcome these obstacles, this study proposes learning-based counterfactual explanations for recommendation(LCER) to provide counterfactual explanations based on personalized recommendations by jointly modeling the factual and counterfactual preference. To achieve consistency between the computation of importance and generation of counterfactual explanations, the proposed LCER endows an optimizable importance for each interacted item, which is supervised by the goal of counterfactual explanations to guarantee its credibility. Because of the model's flexibility, the trade-off between quality and length can be customized by setting different proportions. The experimental results on four real-world datasets demonstrate the effectiveness of the proposed LCER over several

关键词： recommender system explainable recommendation item-based explanation counterfactual inference counterfactual explanation

来源：

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Monochromatic energy computed tomography image for active intestinal hemorrhage:A model investigation

引用

World Journal of Gastroenterology 2015年第1期21卷 214-220页

作者： wen-dong liu Xing-Wang Wu Jun-Mei Hu Bin Wang Bin liu Department of Radiology the First Affiliated Hospital of Anhui Medical University Department of Epidemiology and Statistics Anhui Medical University

AIM: To investigate the value of computed tomography(CT) spectral imaging in the evaluation of intestinal ***: Seven blood flow rates were simulated in *** spectral CT and mixed-energy CT scanswere performed for each rate(0.5,0.4,0.3,0.2,0.1,0.05 and 0.025 m L/min).The detection rates and the contrast-to-noise ratios(CNRs) of the contrast agent extravasation regions were compared between the two scanning methods in the arterial phase(AP) and the portal venous phase(PVP).Comparisons of the CNR values between the PVP and the AP were made for each energy level and carried out using a completely random t test.A χ2 test was used to compare the detection rates obtained from the two scanning ***: The total detection rates for energy spectral CT and mixed-energy CT in the AP were 88.57%(31/35) and 65.71%(23/35),respectively,and the difference was significant(χ2 = 5.185,P = 0.023); the total detection rates in the PVP were 100.00%(35/35) and 91.4%(32/35),respectively,and the difference was not significant(χ2 = 1.393,P = 0.238).In the AP,the CNR of the contrast agent extravasation regions was 3.58 ± 2.09 on the mixed-energy CT images,but the CNRs were 8.78 ± 7.21 and 8.83 ± 6.75 at 50 and 60 keV,respectively,on the single-energy CT images,which were significantly different(3.58 ± 2.09 vs 8.78 ± 7.21,P = 0.031; 3.58 ± 2.09 vs 8.83 ± 6.75,P = 0.029).In the PVP,the differences between the CNRs at 40,50 and 60 keV different monochromatic energy levels and the polychromatic energy images were significant(19.35 ± 10.89 vs 11.68 ± 6.38,P = 0.010; 20.82 ± 11.26 vs 11.68 ± 6.38,P = 0.001; 20.63 ± 10.07 vs 11.68 ± 6.38,P = 0.001).The CNRs at the different energy levels in the AP and the PVP were significantly different(t =-2.415,-2.380,-2.575,-2.762,-2.945,-3.157,-3.996 and-3.189).CONCLUSION: Monochromatic energy imaging spectral CT is superior to polychromatic energy images for the detection of intestinal hemorrhage,and the detection was easier in t

关键词： Spectral imaging Computed tomography Monochromatic

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Achieving strength-ductility synergy in novel paramagnetic Fe-based medium-entropy alloys through deep cryogenic deformation

引用

Rare Metals 2024年第9期43卷 4493-4507页

作者： Hu-wen Ma Yan-Chun Zhao Li Feng Tian-Zeng liu Zhi-Qi Yu Bo Jin Wang-Chun Duan Peter K.Liaw dong Ma State Key Laboratory of Advanced Processing and Recycling of Non-Ferrous Metals Lanzhou University of TechnologyLanzhou 730050China wenzhou Pump and Valve Engineering Research Institute Lanzhou University of TechnologyWenzhou 325105China Iron and Steel Research Institute Jiuquan Iron and Steel CorporationJiayuguan 735100China Lanzhou Resources and Environment Voc-Tech University Lanzhou 730123China Department of Materials Science and Engineering The University of TennesseeKnoxvilleTN 37996-2200USA Neutron Science Center Songshan Lake Materials LaboratoryDongguan 523808China

Cryogenic pre-deformation treatment has been widely used to effectively improve the comprehensive mechanical properties of steels and novel ***,the dislocation evolution and phase transformation induced by different degrees of deep cryogenic deformation are not yet fully *** this study,the effects of multiple cryogenic pre-treatments on the mechanical properties and deformation mechanisms of a paramagnetic Fe_(63.3)Mn_(14-)Si_(9.1)Cr_(9.8)C_(3.8)medium-entropy alloy(MEA)were investigated,leading to the discovery of a pretreated MEA that exhibits exceptional mechanical properties,including a fracture strength of 3.0 GPa,plastic strain of 26.1%and work-hardening index of *** addition,X-ray diffraction(XRD)and transmission electron microscopy(TEM)analyses revealed that multiple cryogenic pre-deformation treatments significantly increased the dislocation density of the MEA(from 9×10^(15)to 4×10^(16)m^(-2)after three pretreatments),along with a transition in the dislocation type from predominantly edge dislocations to mixed dislocations(including screw-and edge-type dislocations).Notably,this pretreated MEA retained its paramagnetic properties(μ_(r)<1.0200)even after *** calculations showed that cryogenic pretreatment can significantly reduce the stacking fault energy of the MEA by a factor of approximately four(i.e.,from 9.7 to2.6 m J·m^(-2)),thereby activating the synergistic effects of transformation-induced plasticity,twinning-induced plasticity and dislocation strengthening *** synergistic effects lead to simultaneous strength and ductility enhancement of the MEA.

关键词： Deep cryogenic transformation Iron-based medium-entropy alloys Dislocation evolution Phase transformation Stacking fault energy

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Peracetic acid catalytic activation using Co3O4 nanocluster decorated graphitic carbon nitride for cefapirin degradation in water

引用

Science China Technological Sciences 2024年第10期67卷 3103-3115页

作者： YIN ShiMei GAO YiXuan CHEN Long LI Fan liu XiaoNa QI Fei liu wen Beijing Key Lab for Source Control Technology of Water Pollution College of Environmental Science and EngineeringBeijing Forestry University Key Laboratory of Water and Sediment Sciences (Ministry of Education) College of Environmental Sciences and EngineeringPeking University School of Environmental Science and Engineering Taiyuan University of Technology State Environmental Protection Key Laboratory of All Material Fluxes in River Ecosystems Peking University Beijing National Laboratory for Molecular Sciences Peking University

This study innovated cobalt-based catalytic nanocomposites（CCNs） for the heterogeneous activation of peracetic acid（PAA）,a green and efficient advanced oxidation process for removing antibiotics from *** Co3O4catalysts exhibit limited activity due to stable electronic *** overcome this,CCNs were synthesized using a one-step condensation method,incorporating a size control strategy that anchors Co3O4nanoclusters within 4-5 nm onto six-fold cavity of g-C3N4,simultaneously modifying the coordination environment of *** alteration significantly enhances the adsorption energy of PAA and catalytic activity compared to commercial *** improved performance of CCN2,in particular,is attributed to a shift in the cobalt d band center towards the Fermi level,facilitating electron delocalization and transfer,enhancing Co-O bonding,and accelerating PAA ***2's superior activity,with a cefapirin（CFP） degradation kinetic rate constant(k1=0.760 min-1)～11 times that of PAA direct oxidation(k1=0.070 min-1) and～5 times that of commercial Co3O4/PAA system(k1=0.164 min-1),is due to its ability to generate dominant reactive species for antibiotic degradation,including alkoxy radicals(CH3C（=O）OO·),high valence metal(CoⅣ（=O）2+),and singlet oxygen（1O2）,with the latter two being crucial for the degradation *** degradation of CFP involved three mechanisms,including direct oxidation by PAA,decarboxylation,and open-ring reactions at nitrogen *** research provides insights into developing high-activity catalysts for efficient PAA activation and antibiotic removal,leveraging the strategy of shifting the transition metal d band center.

关键词： peracetic acid Co3O4 nanocluster high valence metal singlet oxygen cefapirin

来源：

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

DeepCRBP:improved predicting function of circRNA-RBP binding sites with deep feature learning

引用

Frontiers of Computer Science 2024年第2期18卷 235-237页

作者： Zishan XU Linlin SONG Shichao liu wen ZHANG College of Informatics Huazhong Agricultural UniversityWuhan 430070China

1 Introduction Recently,there is growing evidence that it is essential to predict the interactions between circRNAs and RBP binding sites for diagnosing diseases and providing a potential target to treat diseases[1].Many studies have predicted the binding sites of circRNA-RBPs by using deep learning methods based on the sequence information of circRNAs for each ***,many methods focus on diversified view of sequence encoding for improving the performance of the predictive models.

关键词： learning Deep function

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案：

请选择收藏分类：

通借通还

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

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：