检索结果-内蒙古大学图书馆

Temporally identity-aware SSD with attentional LSTM

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Chen, Xingyu Yu, Junzhi Wu, Zhengxing University of Chinese Academy of Sciences Beijing100049 China State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China

Temporal object detection has attracted significant attention, but most popular detection methods can not leverage the rich temporal information in videos. Very recently, many different algorithms have been developed for video detection task, but real-time online approaches are frequently deficient. In this paper, based on attention mechanism and convolutional long short-term memory (ConvLSTM), we propose a temporal signal-shot detector (TSSD) for real-world detection. Distinct from previous methods, we take aim at temporally integrating pyramidal feature hierarchy using ConvLSTM, and design a novel structure including a low-level temporal unit as well as a high-level one (HL-TU) for multi-scale feature maps. Moreover, we develop a creative temporal analysis unit, namely, attentional ConvLSTM (AC-LSTM), in which a temporal attention module is specially tailored for background suppression and scale suppression while a ConvLSTM integrates attention-aware features through time. An association loss is designed for temporal coherence. Besides, online tubelet analysis (OTA) is exploited for identification. Finally, our method is evaluated on ImageNet VID dataset and 2DMOT15 dataset. Extensive comparisons on the detection and tracking capability validate the superiority of the proposed approach. Consequently, the developed TSSD-OTA is fairly faster and achieves an overall competitive performance in terms of detection and tracking. The source code will be made available. Copyright © 2018, The Authors. All rights reserved.

关键词： Object detection

Evolutionary dynamics of cryptocurrency transaction networks: An empirical study

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Liang, Jiaqi Li, Linjing Zeng, Daniel State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China School of Computer and Control Engineering University of Chinese Academy of Sciences Beijing China Department of Management of Information Systems University of Arizona TucsonAZ United States

Cryptocurrency is a well-developed blockchain technology application that is currently a heated topic throughout the world. The public availability of transaction histories offers an opportunity to analyze and compare different cryptocurrencies. In this paper, we present a dynamic network analysis of three representative blockchain-based cryptocurrencies: Bitcoin, Ethereum, and Namecoin. By analyzing the accumulated network growth, we find that, unlike most other networks, these cryptocurrency networks do not always densify over time, and they are changing all the time with relatively low node and edge repetition ratios. Therefore, we then construct separate networks on a monthly basis, trace the changes of typical network characteristics (including degree distribution, degree assortativity, clustering coefficient, and the largest connected component) over time, and compare the three. We find that the degree distribution of these monthly transaction networks cannot be well fitted by the famous power-law distribution, at the same time, different currency still has different network properties, e.g., both Bitcoin and Ethereum networks are heavy-tailed with disassortative mixing, however, only the former can be treated as a small world. These network properties reflect the evolutionary characteristics and competitive power of these three cryptocurrencies and provide a foundation for future research. Copyright © 2018, The Authors. All rights reserved.

关键词： Blockchain

Adaptive and azimuth-aware fusion network of multimodal local features for 3D object detection

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Tian, Yonglin Wang, Kunfeng Wang, Yuang Tian, Yulin Wang, Zilei Wang, Fei-Yue Department of Automation University of Science and Technology of China Hefei230027 China State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China College of Information Science and Technology Beijing University of Chemical Technology Beijing100029 China University of Science and Technology Beijing Beijing100083 China North China University of Technology Beijing100144 China

This paper focuses on the construction of stronger local features and the effective fusion of image and LiDAR data. We adopt different modalities of LiDAR data to generate richer features and present an adaptive and azimuth-aware network to aggregate local features from image, bird's eye view maps and point cloud. Our network mainly consists of three subnetworks: ground plane estimation network, region proposal network and adaptive fusion network. The ground plane estimation network extracts features of point cloud and predicts the parameters of a plane which are used for generating abundant 3D anchors. The region proposal network generates features of image and bird's eye view maps to output region proposals. To integrate heterogeneous image and point cloud features, the adaptive fusion network explicitly adjusts the intensity of multiple local features and achieves the orientation consistency between image and LiDAR data by introduce an azimuth-aware fusion module. Experiments are conducted on KITTI dataset and the results validate the advantages of our aggregation of multimodal local features and the adaptive fusion network. Copyright © 2019, The Authors. All rights reserved.

关键词： Object recognition

Discrete-time Optimal Zero-sum Games for Nonlinear systems via Adaptive Dynamic Programming 6

学校读者我要写书评

暂无评论

Discrete-time Optimal Zero-sum Games for Nonlinear Systems v...

2017 IEEE 6th Data Driven control and Learning systems Conference (DDCLS’17)

作者： Qinglai Wei Ruizhuo Song Yancai Xu Derong Liu Qiao Lin The State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences School of Automation and Electrical Engineering University of Science and Technology Beijing

ISBN: (纸本)9781509054626

In this paper, a novel discrete-time iterative zero-sum adaptive dynamic programming(ADP) algorithm is developed for solving the optimal control problems of nonlinear systems. Two iteration processes, which are lower and upper iterations, are employed to solve the lower and upper value functions, respectively. Arbitrary positive semi-definite functions are acceptable to initialize the upper and lower iterations of the iterative zero-sum ADP algorithm. It is proven that the upper and lower value functions converge to the optimal performance index function if the optimal performance index function exists, where the existence criterion of the optimal performance index function is unnecessary. Simulation examples are given to illustrate the effective performance of the present method.

关键词： Approximate dynamic programming adaptive dynamic programming zero-sum game optimal control

Operation conditions design based on the qualitative and quantitative analysis for glycosylation reactions 36

学校读者我要写书评

暂无评论

Operation conditions design based on the qualitative and qua...

第36届中国控制会议

作者： Yuhan Nan Jing Wang Jinlin Zhou Long Cheng College of Information Science and Technology Beijing University of Chemical Technology State key laboratory of management and control for complex systems Institute of AutomationChinese Academy of Sciences

ISBN: (纸本)9781538629185

Glycosylation reactions play a significant role on the targeted therapy of monoclonal antibody *** this paper,a new method implemented by combination of qualitative analysis and quantitative analysis is introduced to optimize the operating conditions of glycosylation reaction ***,gain matrix based on ANOVA design is presented to evaluate the relationship between reaction output(different glycan classes) and system input(glycosylation enzymes).Due to the high dimension and high nonlinearity of glycosylation reactions,the singular value decomposition of process gain matrix is adopted to qualitatively analyze the effect of glycosylation enzymes in order to find the appropriate operation ***,genetic algorithm is used to quantitatively optimize the operation conditions for desired target of monoclonal antibody *** quantitative operation conditions are consistent with the reaction mechanism and qualitative analysis results of process gain.

关键词： Operating Conditions Optimization Gain Matrix Genetic Algorithm

A survey of generative adversarial networks

学校读者我要写书评

暂无评论

A survey of generative adversarial networks

Chinese Automation Congress

作者： Kongtao Zhu Xiwei Liu Hongxue Yang Beijing Polytechnic College of Telecom. Eng. Beijing China The State Key Laboratory of Management and Control for Complex Systems Qingdao Academy of Intelligent Industries Qingdao Beijing China

Generative adversarial networks (GANs) coming from the game theory allow machines to learn deep representations without extra training data. By training two adversarial networks, including a generator and a discriminator, GANs could get the distribution of the real samples. This capability makes it a prospect learning method in image synthesis, image recognition, image translation etc. In this paper, we survey the state of the art of GANs by categorizing the GANs into four classifications on the basis of GANs' functions and list two application domains: vision computing & natural language processing (NLP) regarding to GANs' applications.

关键词： Generative adversarial networks Vision computing Natural language processing (NLP)

Reinforcement Learning-Based Predictive control for Autonomous Electrified Vehicles

学校读者我要写书评

暂无评论

Reinforcement Learning-Based Predictive Control for Autonomo...

2018 IEEE Intelligent Vehicles Symposium, IV 2018

作者： Liu, Teng Yang, Chao Hu, Chuanzheng Wang, Hong Li, Li Cao, Dongpu Wang, Fei-Yue State Key Laboratory of Automotive Simulation and Control Jilin University China Vehicle Intelligence Pioneers Inc. Qingdao Shandong266109 China Jiangsu XCMG Construction Machinery Research Institute Ltd. Xuzhou Jiangsu221004 China Mechanical and Mechatronics Engineering Department Waterloo University N2L 3G1 Canada Faculty of Department of Automation Tsinghua University China State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China

ISBN: (纸本)9781538644522

This paper proposes a learning-based predictive control technique for self-driving hybrid electric vehicle (HEV). This approach is a hierarchical framework. The higher-level is a human-like driver model, which is applied to predict accelerations in the car following situation to replicate a human driver's demonstrations. The lower-level is a reinforcement learning (RL)-based controller, which enforces the battery and fuel consumption constraints to improve energy efficiency of HEV. In addition, we present induced matrix norm (IMN) to handle cases that the training data cannot provide sufficient information on how to operate in current driving situation. Simulation results illustrate that the proposed method can reproduce human driver's driving style and promote fuel economy. © 2018 IEEE.

关键词： Reinforcement learning

A Kolmogorov-Smirnov Test to Detect Changes in Stationarity in Big Data

学校读者我要写书评

暂无评论

IFAC-PapersOnLine 2017年第1期50卷 14260-14265页

作者： Zhao, Dongbin Bu, Li Alippi, Cesare Wei, Qinglai The State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences University of Chinese Academy of Sciences Beijing100190 China The Politecnico di Milano Università della Svizzera italiana Italy

The paper proposes an effective change detection test for online monitoring data streams by inspecting the least squares density difference (LSDD) features extracted from two non-overlapped windows. The first window contains samples associated with the pre-change probability distribution function (pdf) and the second one with the post-change one (that differs from the former if a change in stationarity occurs). This method can detect changes by also controlling the false positive rate. However, since the window sizes is fixed after the test has been configured (it has to be small to reduce the execution time), the method may fail to detect changes with small magnitude which need more samples to reach the requested level of confidence. In this paper, we extend our work to the Big Data framework by applying the Kolmogorov-Smirnov test (KS test) to infer changes. Experiments show that the proposed method is effective in detecting changes. © 2017

关键词： Testing Big data Computational complexity Data streams Distribution functions Change detection tests Data framework Density difference False positive rates K S test Kolmogorov Smirnov test LSDD Online monitoring

A right-of-way based strategy to implement safe and efficient driving at non-signalized intersections for automated vehicles

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Xing, Yadong Zhao, Can Li, Zhiheng Zhang, Yi Li, Li Wang, Fei-Yue Wang, Xiao Wang, Yujing Su, Yuelong Cao, Dongpu Department of Automation Tsinghua University Beijing100084 China Graduate School at Shenzhen Tsinghua University Shenzhen518055 China Department of Automation BNRist Tsinghua University Beijing100084 China State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China Traffic Management Solution Division AutoNavi Software Co Beijing100102 China Driver Cognition and Automated Driving Laboratory University of Waterloo WaterlooN2L 3G1 Canada

Non-signalized intersection is a typical and common scenario for connected and automated vehicles (CAVs). How to balance safety and efficiency remains difficult for researchers. To improve the original Responsibility Sensitive Safety (RSS) driving strategy on the non-signalized intersection, we propose a new strategy in this paper, based on right-of-way assignment (RWA). The performances of RSS strategy, cooperative driving strategy, and RWA based strategy are tested and compared. Testing results indicate that our strategy yields better traffic efficiency than RSS strategy, but not satisfying as the cooperative driving strategy due to the limited range of communication and the lack of long-term planning. However, our new strategy requires much fewer communication costs among vehicles. Copyright © 2019, The Authors. All rights reserved.

关键词： Vehicles