检索结果-南通市图书馆

TSC prediction and dynamic control of BOF steelmaking with state-of-the-art machine learning and deep learning methods

在线全文

学校读者我要写书评

暂无评论

Journal of Iron and Steel Research(International) 2024年第1期31卷 174-194页

作者： tian-yi Xie Cai-dong Zhang Quan-lin Zhou zhi-qiang tian Shuai Liu Han-jie Guo Material Technology Research Institute Hesteel GroupShijiazhuang050023HebeiChina Tangsteel Company Hesteel GroupTangshan063611HebeiChina School of Artificial Intelligence Beijing Technology and Business UniversityBeijing100048China School of Metallurgical and Ecological Engineering University of Science and Technology BeijingBeijing100083China

Mathematical(data-driven)models based on state-of-the-art(SOTA)machine learning and deep learning models and data collected from 12,786 heats were established to predict the values of temperature,sample,and carbon(TSC)test,including temperature of molten steel(TSC-Temp),carbon content(TSC-C)and phosphorus content(TSC-P),which made prepa-ration for eliminating the TSC test.To maximize the prediction accuracy of the proposed approach,various models with different inputs were implemented and compared,and the best models were applied to the production process of a Hesteel Group steelmaking plant in China in the field.The number of tabular features(hot metal information,scrap,additives,blowing practices,and preset values)was expanded,and time series(off-gas profiles and blowing practice curves)that could reflect the entire steelmaking process were introduced as inputs.First,the latest machine learning models(LightGBM,CatBoost,TabNet,and NODE)were used to make predictions with tabular features,and the best coefficient of determination R^(2)values obtained for TSC-P,TSC-C and TSC-Temp predictions were 0.435(LightGBM),0.857(Cat-Boost)and 0.678(LightGBM),respectively,which were higher than those of classic models(backpropagation and support vector machine).Then,making predictions was performed by using SOTA time series regression models(SCINet,DLinear,Informer,and MLSTM-FCN)with original time series,SOTA image regression models(NesT,CaiT,ResNeXt,and GoogLeNet)with resized time series,and the proposed Concatenate-Model and Parallel-Model with both tabular features and time series.Through optimization and comparisons,it was finally determined that the Concatenate-Model with MLSTM-FCN,SCINet and Informer as feature extractors performed the best,and its R^(2)values for predicting TSC-P,TSC-C and TSC-Temp reached 0.470,0.858 and 0.710,respectively.Its field test accuracies for TSC-P,TSC-C and TSC-Temp were 0.459,0.850 and 0.685,respectively.A related importance analysis was carried out,and dynamic control methods based on pred

关键词： BOF steelmaking In-blow prediction TSC test Machine learning Deep learning Field application

Sulfur doped iron-nitrogen-hard carbon nanosheets as efficient and robust noble metal-free catalysts for oxygen reduction reaction in PEMFC

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

Journal of Energy Chemistry 2024年第2期89卷 422-433,I0010页

作者： Bin Liu Jiawang Li Bowen Yan Qi Wei Xingyu Wen Huarui Xie Huan He Pei Kang Shen zhi Qun tian Collaborative Innovation Center of Sustainable Energy Materials School of Physical Science and TechnologyGuangxi UniversityGuangxi Key Laboratory of Electrochemical Energy MaterialsState Key Laboratory of Featured Metal Materials and Life-cycle Safety for Composite StructuresNanning 530004GuangxiChina

Transition metal-nitrogen-carbon(M-N-C)as a promising substitute for the conventional noble metalbased catalyst still suffers from low activity and durability for oxygen reduction reaction(ORR)in proton exchange membrane fuel cells(PEMFCs).To tackle the issue,herein,a new type of sulfur-doped ironnitrogen-hard carbon(S-Fe-N-HC)nanosheets with high activity and durability in acid media were developed by using a newly synthesized precursor of amide-based polymer with Fe ions based on copolymerizing two monomers of 2,5-thiophene dicarboxylic acid(TDA)as S source and 1,8-diaminonaphthalene(DAN)as N source via an amination reaction.The as-synthesized S-Fe-N-HC features highly dispersed atomic Fe Nxmoieties embedded into rich thiophene-S doped hard carbon nanosheets filled with highly twisted graphite-like microcrystals,which is distinguished from the majority of M-N-C with soft or graphitic carbon structures.These unique characteristics endow S-Fe-N-HC with high ORR activity and outstanding durability in 0.5 M H_(2)SO_(4).Its initial half-wave potential is 0.80 V and the corresponding loss is only 21 m V after 30,000 cycles.Meanwhile,its practical PEMFC performance is a maximum power output of 628.0 mW cm^(-2)and a slight power density loss is 83.0 m W cm^(-2)after 200-cycle practical operation.Additionally,theoretical calculation shows that the activity of Fe Nxmoieties on ORR can be further enhanced by sulfur doping at meta-site near FeN_(4)C.These results evidently demonstrate that the dual effect of hard carbon substrate and S doping derived from the precursor platform of amid-polymers can effectively enhance the activity and durability of Fe-N-C catalysts,providing a new guidance for developing advanced M-N-C catalysts for ORR.

关键词： Transition metal-nitrogen-carbon Oxygen reduction reaction Hard carbon Amide based polymer reaction Proton exchange membrane cells

Link-Privacy Preserving Graph Embedding Data Publication with Adversarial Learning

在线全文

学校读者我要写书评

暂无评论

Tsinghua Science and Technology 2022年第2期27卷 244-256页

作者： Kainan Zhang zhi tian zhipeng Cai Daehee Seo the Department of Computer Science Georgia State UniversityAtlantaGA 30303USA the Department of Electrical and Computer Engineering George Mason UniversityFairfaxVA 22030USA the National Center of Excellence in Software Sangmyung UniversitySeoul 03016Republic of Korea

The inefficient utilization of ubiquitous graph data with combinatorial structures necessitates graph embedding methods,aiming at learning a continuous vector space for the graph,which is amenable to be adopted in traditional machine learning algorithms in favor of vector representations.Graph embedding methods build an important bridge between social network analysis and data analytics,as social networks naturally generate an unprecedented volume of graph data continuously.Publishing social network data not only brings benefit for public health,disaster response,commercial promotion,and many other applications,but also gives birth to threats that jeopardize each individual’s privacy and security.Unfortunately,most existing works in publishing social graph embedding data only focus on preserving social graph structure with less attention paid to the privacy issues inherited from social networks.To be specific,attackers can infer the presence of a sensitive relationship between two individuals by training a predictive model with the exposed social network embedding.In this paper,we propose a novel link-privacy preserved graph embedding framework using adversarial learning,which can reduce adversary’s prediction accuracy on sensitive links,while persevering sufficient non-sensitive information,such as graph topology and node attributes in graph embedding.Extensive experiments are conducted to evaluate the proposed framework using ground truth social network datasets.

关键词： graph embedding privacy preservation adversarial learning

Design space exploration of neural network accelerator based on transfer learning

在线全文

学校读者我要写书评

暂无评论

High Technology Letters 2023年第4期29卷 416-426页

作者：吴豫章 zhi tian SONG Xinkai LI Xi School of Computer Science University of Science and Technology of ChinaHefei 230027P.R.China State Key Lab of Processors Institute of Computing TechnologyChinese Academy of SciencesBeijing 100190P.R.China

With the increasing demand of computational power in artificial intelligence(AI)algorithms,dedicated accelerators have become a necessity.However,the complexity of hardware architectures,vast design search space,and complex tasks of accelerators have posed significant challenges.Tra-ditional search methods can become prohibitively slow if the search space continues to be expanded.A design space exploration(DSE)method is proposed based on transfer learning,which reduces the time for repeated training and uses multi-task models for different tasks on the same processor.The proposed method accurately predicts the latency and energy consumption associated with neural net-work accelerator design parameters,enabling faster identification of optimal outcomes compared with traditional methods.And compared with other DSE methods by using multilayer perceptron(MLP),the required training time is shorter.Comparative experiments with other methods demonstrate that the proposed method improves the efficiency of DSE without compromising the accuracy of the re-sults.

关键词： design space exploration(DSE) transfer learning neural network accelerator multi-task learning

Heterophase junction engineering-induced Co spin-state modulation of CoSe2 for large-current hydrogen evolution reaction

在线全文

学校读者我要写书评

暂无评论

Rare Metals 2024年第6期43卷 2660-2670页

作者： Bao-Chai Xu Ya-Ping Miao Min-Qin Mao Dong-Lian Li Song Xie Wei-Hong Jin Shu Xiao Jing Wen Zaenab Abd-Allah zhi-tian Liu Xiang Peng Paul K.Chu Hubei Key Laboratory of Plasma Chemistry and Advanced Materials School of Materials Science and EngineeringWuhan Institute of Technology School of Textile Science and Engineering Xi'an Polytechnic University Institute of Advanced Wear and Corrosion Resistant and Functional Materials Jinan University Institute of Safety Science and Engineering South China University of Technology Department of Engineering Manchester Metropolitan University Department of Physics Department of Materials Science and Engineeringand Department of Biomedical EngineeringCity University of Hong Kong

Efficient electrocatalysts are vital to large-current hydrogen production in commercial water splitting for green energy generation.Herein,a novel heterophase engineering strategy is described to produce polymorphic CoSe2electrocatalysts.The composition of the electrocatalysts consisting of both cubic CoSe2（c-CoSe2） and orthorhombic CoSe2（o-CoSe2） phases can be controlled precisely.Our results demonstrate that junction-induced spin-state modulation of Co atoms enhances the adsorption of intermediates and accelerates charge transfer resulting in superior large-current hydrogen evolution reaction（HER）properties.Specifically,the CoSe2-based heterophase catalyst with the optimal c-CoSe2content requires an overpotential of merely 240 mV to achieve 1,000 mA·cm-2as well as a Tafel slope of 50.4 mV·dec-1.Furthermore,the electrocatalyst can maintain a large current density of 1,500 mA·cm-2for over 320 h without decay.The results reveal the advantages and potential of heterophase junction engineering pertaining to design and fabrication of low-cost transition metal catalysts for large-current water splitting.

关键词： Heterophase junction Spin-state Hydrogen evolution reaction Large current electrochemical hydrogen production Water splitting

政务元宇宙赋能公共治理:治理契机与风险因应

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

华侨大学学报（哲学社会科学版） 2024年第3期 76-87页

作者：金华祝天智江苏师范大学公共管理与社会学院江苏徐州221116

政务元宇宙是虚拟世界与现实世界虚实相生的政府治理新模式和公共服务供给新体系。凭借其超强大技术群支撑,政务元宇宙在公共治理领域不仅能够表现出强大的组织整合功能、任务激励功能及社区自治功能,还能驱动组织内部变革与外在治理的... 详细信息

政务元宇宙是虚拟世界与现实世界虚实相生的政府治理新模式和公共服务供给新体系。凭借其超强大技术群支撑,政务元宇宙在公共治理领域不仅能够表现出强大的组织整合功能、任务激励功能及社区自治功能,还能驱动组织内部变革与外在治理的双重转型。从组织内部变革来看,政务元宇宙将推动政府服务效能优化、政民互动更加高效及治理规则不断创新。从外在治理转型来看,政务元宇宙将实现公共治理领域信用体系重塑、安全体系重构、价值体系重建与治理体系精细化。随着虚拟世界对现实世界影响力的不断增强,传统科层制度与治理范式将遭遇不同程度的挑战,如数字技术依赖引发政治安全、数字人身份引发虚拟世界监管难题、数字世界新规使得传统科层规则悬置。国家须积极布局政务元宇宙,通过推动关键数字资产国有化及算力设施国家化以打破数字技术外部依赖;以国家信用为背书加快构建数字人身份监管新规定;增加国家在元宇宙空间的话语权与行动能力以打破科技巨头规则垄断。

关键词：政务元宇宙治理契机风险因应

Experimental research on spinal metastasis with mouse models

在线全文

学校读者我要写书评

暂无评论

Chinese Medical Journal 2023年第24期136卷 3008-3009页

作者： Kun Zhang Yi Feng Xiaochen Qiao Yang Yu Zelong Song Zhuohao Liu zhi tian Song Chen Xuesong Zhang Xiangyu Wang Second Clinical Medical College Shanxi Medical UniversityTaiyuanShanxi 030001China Department of Orthopedics The Second Hospital of Shanxi Medical UniversityShanxi Key laboratory of Bone and Soft Tissue Injury RepairTaiyuanShanxi 030001China Department of Orthopedics JinZhong Hospital Affiliated to Shanxi Medical UniversityJinzhongShanxi 030600China Nankai University School of Medicine Nankai UniversityTianjin 300071China Department of Orthopaedics the PLA General HospitalBeijing 100853China Department of Pain medicine First Medical Centerthe PLA General HospitalBeijing 100853China

To the Editor:The spine is a common site for tumor metastasis.Patients with spinal metastases have pain,pathological fractures and spinal deformities owing to the tumor invasion of spinal bones.At present,the mouse model is a very important tool for spinal metastases-related studies.[1]It is crucial to elucidate the pathophysiological mechanism underlying spinal metastases,improve the diagnosis and treatment guidelines,and develop new therapeutic methods.In this study,we have summarized various mouse spinal metastasis models and the applicable research contents.

关键词： metastasis invasion diagnosis

LACC:a hardware and software co-design accelerator for deep neural networks

在线全文

学校读者我要写书评

暂无评论

High Technology Letters 2021年第1期27卷 62-67页

作者：于涌 zhi tian Zhou Shengyuan State Key Laboratory of Computer Architecture Institute of Computing TechnologyChinese Academy of SciencesBeijing 100190P.R.China School of Computer Science and Technology University of Chinese Academy of SciencesBeijing 100049P.R.China Cambricon Technologies Ltd Beijing 100190P.R.China

With the increasing of data size and model size,deep neural networks(DNNs)show outstanding performance in many artificial intelligence(AI)applications.But the big model size makes it a challenge for high-performance and low-power running DNN on processors,such as central processing unit(CPU),graphics processing unit(GPU),and tensor processing unit(TPU).This paper proposes a LOGNN data representation of 8 bits and a hardware and software co-design deep neural network accelerator LACC to meet the challenge.LOGNN data representation replaces multiply operations to add and shift operations in running DNN.LACC accelerator achieves higher efficiency than the state-of-the-art DNN accelerators by domain specific arithmetic computing units.Finally,LACC speeds up the performance per watt by 1.5 times,compared to the state-of-the-art DNN accelerators on average.

关键词： deep neural network(DNN) domain specific accelerator domain specific data type

一种基于自适应PoT量化的无乘法神经网络训练方法

在线全文

学校读者我要写书评

暂无评论

高技术通讯 2024年第6期34卷 567-577页

作者：刘畅张蕊支天中国科学院大学北京100049 中国科学院计算技术研究所智能处理器研究中心北京100190 中国科学院计算技术研究所处理器芯片全国重点实验室北京100190

当前的深度神经网络的训练过程中需要包含大量的全精度乘累加(MAC)操作,导致神经网络模型的线性层(包含卷积层和全连接层)的计算过程所需的能耗占整体能耗的绝大部分,达90%以上。本文提出了一种自适应逐层缩放的量化训练方法,可支持在... 详细信息

当前的深度神经网络的训练过程中需要包含大量的全精度乘累加(MAC)操作,导致神经网络模型的线性层(包含卷积层和全连接层)的计算过程所需的能耗占整体能耗的绝大部分,达90%以上。本文提出了一种自适应逐层缩放的量化训练方法,可支持在神经网络计算全流程(前向传播和后向传播)将全部线性层中的全精度乘法替换为4位定点数加法计算和1位异或运算。实验结果表明,上述方法在能耗和准确率方面都优于现有方法,可支撑在训练过程中减少达95.8%的线性层能耗,在ImageNet数据集上的卷积神经网络和在WMT En-De任务上的Transformer网络得到小于1%的精度损失。

关键词：神经网络量化训练加速低能耗