检索结果-内蒙古大学图书馆

Answer Extraction with Graph Attention Network for Knowledge Graph Question Answering

学校读者我要写书评

暂无评论

Answer Extraction with Graph Attention Network for Knowledge...

International Conference on Computer and Communications (ICCC)

作者： Jie Zhang Zhongmin Pei Wei Xiong Zhangkai Luo Science and Technology on Complex Electronic System Simulation Laboratory Space Engineering University Beijing China

ISBN: (数字)9781728186351

ISBN: (纸本)9781728186368

In the knowledge graph question answering, the graph neural network can be used to encode the subgraph nodes related to the question entity to select the correct answer node. However, existing researches mainly focus on the modalities for the node encoding with graph neural network, ignoring that different types of subgraphs have different requirements for encoding information. To overcome the problem, this paper divides the subgraph into two types: the searching graph and the extending graph. Then we propose an answer extraction method with graph attention network for the searching graph, which can weight the information of neighbor nodes with different attention instead of the average. The hierarchical attention is also introduced to integrate question information into the subgraph node embedding to obtain the node presentation with question dependency. The accuracy of 48.2% is achieved on the CommonsenseQA dataset, which is much higher than the random guess (20%). In addition, the accuracy of the simplified model with no hierarchical attention decreases by 3.5%, which indicates the hierarchical attention mechanism can improve the predictive performance of the proposed model.

关键词： Computational modeling Predictive models Knowledge discovery Search problems Graph neural networks Encoding Task analysis

Synergistic Interface Effects in Composite Dielectrics: Insights into Charge Trapping Regulation through Multiscale Modeling

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhao, Haoxiang An, Lixuan Zhang, Daning Yang, Xiong Yao, Huanmin Zhang, Guanjun Mu, Haibao Baumeier, Björn State Key Laboratory of Electrical Insulation and Power Equipment School of Electrical Engineering Xi'an Jiaotong University Xi'An710049 China Department of Mathematics and Computer Science Eindhoven University of Technology Eindhoven5600MB Netherlands Department of Applied Physics Eindhoven University of Technology Eindhoven5600MB Netherlands KERMIT Department of Data Analysis and Mathematical Modelling Ghent University Ghent9000 Belgium Institute for Complex Molecular Systems Eindhoven University of Technology Eindhoven5600MB Netherlands

The rapid development of modern energy applications drives an urgent need to enhance the dielectric strength of energy storage dielectrics for higher power density. Interface design is a promising strategy to regulate the crucial charge transport process determining dielectric strength. However, the targeted exploitation of interface effects on charge transport is limited due to a lack of fundamental understanding of the underlying mechanisms involving elementary electronic processes and details of the intricate interplay of characteristics of molecular building blocks and the interfacial morphology - details that cannot fully be resolved with experimental methods. Here we employ a multiscale modeling approach linking the quantum properties of the charge carriers with nano- and mesoscale structural details of complex interfaces. Applied to a prototypical application-proven cellulose-oil composite with interfaces formed between oil, disordered, and crystalline cellulose regions, this approach demonstrates that charges are trapped in the disordered region. Specifically, it unveils this trapping as a synergistic effect of two transport-regulating interface mechanisms: back-transfer to the oil region is suppressed by energetic factors, while forward-transfer to the crystalline cellulose is suppressed by low electronic coupling. The insight into the molecular origins of interface effects via dual-interface regulation offers new development paths for advanced energy materials. © 2024, CC BY.

关键词： Nanocrystals

Discretizing Continuous Action Space with Unimodal Probability Distributions for On-Policy Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhu, Yuanyang Wang, Zhi Zhu, Yuanheng Chen, Chunlin Zhao, Dongbin The Department of Control Science and Intelligent Engineering School of Management and Engineering Nanjing University Nanjing210093 China The Research Center for Novel Technology of Intelligent Equipment Nanjing University Nanjing210093 China The State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China The School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China

For on-policy reinforcement learning, discretizing action space for continuous control can easily express multiple modes and is straightforward to optimize. However, without considering the inherent ordering between the discrete atomic actions, the explosion in the number of discrete actions can possess undesired properties and induce a higher variance for the policy gradient estimator. In this paper, we introduce a straightforward architecture that addresses this issue by constraining the discrete policy to be unimodal using Poisson probability distributions. This unimodal architecture can better leverage the continuity in the underlying continuous action space using explicit unimodal probability distributions. We conduct extensive experiments to show that the discrete policy with the unimodal probability distribution provides significantly faster convergence and higher performance for on-policy reinforcement learning algorithms in challenging control tasks, especially in highly complex tasks such as Humanoid. We provide theoretical analysis on the variance of the policy gradient estimator, which suggests that our attentively designed unimodal discrete policy can retain a lower variance and yield a stable learning process. © 2024, CC BY.

关键词： Contrastive Learning

The Multi-Blockchain Based Data Trading Markets with Novel Pricing Mechanisms

学校读者我要写书评

暂无评论

SSRN

SSRN 2022年

作者： Li, Juanjuan Li, Junqing Wang, Xiao Qin, Rui Yuan, Yong Wang, Fei-Yue The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China College of Computer Science and Engineering Shandong University of Science and Technology Qingdao266590 China Qingdao Academy of Intelligent Industries Qingdao266109 China School of Mathematics Renmin University of China Beijing100872 China Macao Institute of Systems Engineering Macau University of Science and Technology Macau 999078 China

Aiming to deal with the challenge of data security and data pricing in data trading markets, we propose the novel multi-blockchain-based framework and design the related pricing mechanisms. Specifically, data recording and trading are conducted separately in the Data Blockchain (DChain) and the Value Blockchain (VChain). Based on these blockchains, two-layer data trading markets are established to manage the initial data trading in the primary market and the data resale in the secondary market. Pricing mechanisms are proposed to protect the formulated data trading markets against strategic trading behaviors and balance the payoffs of both suppliers and users. In the regular data trading on VChain-S2D, two auction models are used to deal with users' strategic bidding in terms of the demand scale, where the incentive-compatible Vickrey-Clarke-Groves (VCG) model is dedicated to the low-demand trading scenario, while the nearly incentive-compatible Monopolistic Price (MP) model is for the high-demand trading scenario. In the temporary data trading on VChain-D2S, a reverse auction mechanism named Two-Stage Obscure Selection (TSOS) is designed to regulate both suppliers' quoting and users' valuation strategies. Moreover, experiments are carried out to validate our theoretical analysis. Our work is expected to enhance the data security and improve the trading efficiency. © 2022, The Authors. All rights reserved.

关键词： Blockchain

Co-N-C Axially Coordination Regulated H 2 O 2 Selectivity Via Water Medicated Recombination of Solute Oh•: A New Route

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Wang, Jing Chen, Junpeng Li, Shuang Zhao, Yuan Zhao, Jiachao Liang, Xinyu Xia, Meirong Yang, Jun Yong, Xue Tse, John S. Key Laboratory of Applied Chemistry China Hebei Key Laboratory of Heavy Metal Deep-Remediation in Water and Resource Reuse School of Environmental and Chemistry Engineering State Key Laboratory of Metastable Materials Science and Technology Yanshan University Qinhuangdao066004 China State Key Laboratory of Multiphase Complex Systems Institute of Process Engineering Chinese Academy of Sciences Beijing100190 China Department of Electrical Engineering and Electronics University of Liverpool United Kingdom Department of Physics University of Saskatchewan SaskatoonSKS7N 5B2 Canada

The cobalt, earth abundant transition metal, embedded in nitrogen doped carbon material as single atom site (Co-N-C) has been manifested as promising electrochemical oxygen reduction reaction (ORR) catalyst, however the unsatisfying production selectivity has hampered its widespread applications. Herein, the H2O2 selectivity of Co-N-C catalyst has been tailored with Co axial functional groups.. Thermodynamically, the selectivity is regulated due to the fine-tunning of the adsorption of the key reaction intermediates (ΔG*OOH), and five functional groups, including -O, -OH, -CN, CH3 and -SO3, endow the Co-N-C catalyst with superior H2O2 selectivity. Importantly, we unravel a new water medicated recombination of solute OH• reaction pathway for H2O2 production, which was the result of dissociation of *HOOH in explicit water environment That is two ·OH species reaction in the liquid environment which originated from the creaking of *OOH intermediates due to the weakened O-O bond by the interaction with surrounding water. This study provides foundational understanding for the ORR catalytic mechanism at the electrochemical interface and open up new avenues for rational design of targeted high efficiency electrocatalysts. © 2024, The Authors. All rights reserved.

关键词： Electrolytic reduction

A Novel Two-Stage Framework for 2D/3D Registration in Neurological Interventions

学校读者我要写书评

暂无评论

A Novel Two-Stage Framework for 2D/3D Registration in Neurol...

IEEE International Conference on Robotics and Biomimetics

作者： De-Xing Huang Xiao-Hu Zhou Xiao-Liang Xie Shi-Qi Liu Zhen-Qiu Feng Jian-Long Hao Zeng-Guang Hou Ning Ma Long Yan The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China The School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China The School of Information Shanxi University of Finance and Economics Taiyuan China The CAS Center for Excellence in Brain Science and Intelligence Technology Beijing China The Joint Laboratory of Intelligence Science and Technology Institute of Systems Engineering Macau University of Science and Technology Taipa Macau The Department of Interventional Neuroradiology Beijing Tiantan Hospital Capital Medical University Beijing China

ISBN: (纸本)9781665481106

2D/3D medical image registration of pre-operative volumes and intra-operative images plays an important role in neurological interventions. However, vast space of transformation parameters makes this task incredibly challenging. In this paper, a novel two-stage framework for 2D/3D registration is proposed. It consists of two basic modules: CNN regression for preliminarily estimating transformation parameters, and centroid alignment for further reducing estimation errors to get refined images. Experimental results on two patients show the proposed framework can achieve superior performances to baseline methods with an inference rate of 22 FPS, satisfying real-time requirements (6~12 FPS) in clinical practice. Extensive ablation studies prove the effectiveness of centroid alignment in significantly improving registration accuracies. These promising results indicate the framework has the potential to be integrated into navigation frameworks for neurological interventions, facilitating treatments of physicians.

关键词： Image registration Estimation error Three-dimensional displays Navigation Biomimetics Medical services Real-time systems

Ultra-sensitive vibrational resonance induced by small disturbances

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Li, Shangyuan Wang, Zhongqiu Yang, Jianhua Sanjuán, Miguel A.F. Huang, Shengping Lou, Litai Jiangsu Key Laboratory of Mine Mechanical and Electrical Equipment School of Mechatronic Engineering China University of Mining and Technology Jiangsu Xuzhou221116 China School of Computer Science and Technology China University of Mining and Technology Jiangsu Xuzhou221116 China Nonlinear Dynamics Chaos and Complex Systems Group Departamento de Física Universidad Rey Juan Carlos Tulipán s/n Móstoles Madrid28933 Spain

We have found two kinds of ultra-sensitive vibrational resonance in coupled nonlinear systems. It is particularly worth pointing out that this ultra-sensitive vibrational resonance is a transient behavior caused by transient chaos. Considering long-term response, the system will transform from transient chaos to periodic response. The pattern of vibrational resonance will also transform from ultra-sensitive vibrational resonance to conventional vibrational resonance. This article focuses on the transient ultra-sensitive vibrational resonance phenomenon. It is induced by a small disturbance of the high-frequency excitation and the initial simulation conditions, respectively. The damping coefficient and the coupling strength are the key factors to induce the ultra-sensitive vibrational resonance. By increasing these two parameters, the vibrational resonance pattern can be transformed from an ultra-sensitive vibrational resonance to a conventional vibrational resonance. The reason for different vibrational resonance patterns to occur lies in the state of the system response. The response usually presents transient chaotic behavior when the ultra-sensitive vibrational resonance appears and the plot of the response amplitude versus the controlled parameters shows a highly fractalized pattern. When the response is periodic or doubly-periodic, it usually corresponds to the conventional vibrational resonance. The ultra-sensitive vibrational resonance not only occurs at the excitation frequency, but it also occurs at some more nonlinear frequency components. The ultra-sensitive vibrational resonance as a transient behavior and the transformation of vibrational resonance patterns are new phenomena in coupled nonlinear systems. © 2023, CC BY-NC-ND.

关键词： Nonlinear systems

Anomaly Detection in Digital Twin Model

学校读者我要写书评

暂无评论

Anomaly Detection in Digital Twin Model

Digital Twins and Parallel Intelligence (DTPI), IEEE International Conference on

作者： Yutong Wang Yansong Cao Fei-Yue Wang The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China Macau Institute of Systems Engineering Macau University of Science and Technology Macau China

As the simulation model of a physical system, digital twin has been widely used in many complicated control systems. Providing an effective way to perform simulation, digital twin makes the evaluation, prediction and optimization process cheaper and easier than on physical systems. For smart manufacturing, digital twin achieves high productivity with less operation and maintenance cost. However, the advantages of digital twin in anomaly detection during manufacturing are always neglected. Taking a two-speed transmission system as an example, we generate three kinds of faults on this digital twin model. After extracting features from the system output, we train an anomaly detection model on these features to classify each type of fault. By this digital twin model, we can find the fault in the whole system at the very beginning, and reduce the time and cost of debugging and diagnosing.

关键词： Productivity Costs Digital twin Time series analysis Predictive models Maintenance engineering Feature extraction

EXACT FIRST PASSAGE TIME DISTRIBUTION FOR SECOND-ORDER REACTIONS IN CHEMICAL NETWORKS

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Rao, Changqian Waxman, David Lin, Wei Song, Zhuoyi School of Mathematical Sciences Fudan University Shanghai200433 China Research Institute of Intelligent Complex Systems Fudan University Shanghai200433 China Institute of Science and Technology for Brain-Inspired Intelligence Fudan University Shanghai200433 China Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence Fudan University Ministry of Education China MOE Frontiers Center for Brain Science Fudan University Shanghai200433 China Zhangjiang Fudan International Innovation Center Shanghai China

The first passage time (FPT) is a generic measure that quantifies when a random quantity reaches a specific state. We consider the FTP distribution in nonlinear stochastic biochemical networks, where obtaining exact solutions of the distribution is a challenging problem. Even simple two-particle collisions cause strong nonlinearities that hinder the theoretical determination of the full FPT distribution. Previous research has either focused on analyzing the mean FPT, which provides limited information about a system, or has considered time-consuming stochastic simulations that do not clearly expose causal relationships between parameters and the system’s dynamics. This paper presents the first exact theoretical solution of the full FPT distribution in a broad class of chemical reaction networks involving A + B → C type of second-order reactions. Our exact theoretical method outperforms stochastic simulations, in terms of computational efficiency, and deviates from approximate analytical solutions. Given the prevalence of bimolecular reactions in biochemical systems, our approach has the potential to enhance the understanding of real-world biochemical processes. © 2024, CC BY.

关键词： Stochastic systems