检索结果-内蒙古大学图书馆

An Event Timeline Extraction Method Based on News Corpus 2

学校读者我要写书评

暂无评论

An Event Timeline Extraction Method Based on News Corpus

2017 IEEE 2nd International Conference on Big Data Analysis（ICBDA 2017）

作者： Yaguang Wu Haichun Sun Chungang Yan Department of Computer Science and Technology Tongji University The Key Laboratory of Embedded System and Service Computing Tongji University College of Police Information Engineering People's Public Security University of China

ISBN: (纸本)9781509036202

Event extraction is an important research point in information extraction area, and news event extraction has a greater practical significance. The existing methods of extracting news event, which starts from the time element, is to identify the date sentences hold by Natural Language Processing and extract the event on the date by Text Clustering. However, they only process the news that holds a time-tag and gives up the news that has no accurate date, which easily leads an extraction deviation of significant events and reduces the accuracy of ordering the significant events. In this paper, improvement for this defect is to calculate the similarities between sentences to put part of the non-time-tag sentences into the right date container, thereby the accuracy of ordering the significant event is *** experiments and compared with the existing methods,the accuracy of ordering the significant events in this paper is improved by 14.6%.

关键词： event extraction text information mining sentence similarity timeline

An Improved Convolutional Neural Network Algorithm for Multi-Label Classification

学校读者我要写书评

暂无评论

An Improved Convolutional Neural Network Algorithm for Multi...

International Conference on Audio, Language and Image Processing, ICALIP

作者： Xinsheng Wang Lijun Sun Zhihua Wei Department of Computer Science and Technology Tongji University Shanghai China Research Center of Big Data and Network Security Tongji University Shanghai China Key Laboratory of Embedded System and Service Computing Tongji University Shanghai China

Recent years conventional neural network(CNN) has been applied to different natural language processing(NLP) tasks such as sentence classification, sentence modeling, etc. Some researchers use CNN to do multi-label classification but their work mainly focus on image rather than text. In this paper, we propose an improved CNN via hierarchical dirichlet process(HDP) model to deal with the multi-label classification problem in NLP. We first apply an HDP model to discard some words which are less important semantically. Then we use word embedding methods to transform words to vectors. Finally, we train CNN based on word vectors. Experimental results demonstrate that our method is superior to most traditional multi-label classification methods and TextCNN in terms of performance.

关键词： Feature extraction Training Convolutional neural networks Natural language processing Vocabulary Transforms

Vision-based Semantic Mapping and Localization for Autonomous Indoor Parking

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Huang, Yewei Zhao, Junqiao He, Xudong Zhang, Shaoming Feng, Tiantian Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai Department of Computer Science and Technology School of Electronics and Information Engineering Tongji University Shanghai School of Surveying and Geo-Informatics Tongji University Shanghai China

In this paper, we proposed a novel and practical solution for the real-time indoor localization of autonomous driving in parking lots. High-level landmarks, the parking slots, are extracted and enriched with labels to avoid the aliasing of low-level visual features. We then proposed a robust method for detecting incorrect data associations between parking slots and further extended the optimization framework by dynamically eliminating suboptimal data associations. Visual fiducial markers are introduced to improve the overall precision. As a result, a semantic map of the parking lot can be established fully automatically and robustly. We experimented the performance of real-time localization based on the map using our autonomous driving platform TiEV, and the average accuracy of 0.3m track tracing can be achieved at a speed of 10kph. Copyright © 2018, The Authors. All rights reserved.

关键词： Indoor positioning systems

Implicit cooperative caching based on information popularity for vehicular networks 12

学校读者我要写书评

暂无评论

Implicit cooperative caching based on information popularity...

12th Workshop on Challenged Networks, CHANTS 2017

作者： Li, Yuhong Wang, Ji Lindgren, Anders Su, Xiang Riekki, Jukka State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai China RISE SICS Sweden Luleå University of Technology Sweden Centre for Ubiquitous Computing University of Oulu Finland

ISBN: (纸本)9781450351447

Information-centric networking (ICN) technology is becoming a popular research topic in vehicular networks due to the connectionless and lightweight characteristics of this networking paradigm. Caching plays an essential role in information-centric networks, but current caching techniques for ICN are not ideal for the dynamic and wireless vehicular networks. This paper presents a caching approach for ICN-based vehicular networks that takes into account both the dynamicity of vehicular networks and the popularity of the information being distributed. We introduce an interval metric for selective caching. With this metric and estimates of information popularity and vehicle density, cooperative caching can be realized without exchanging cache management information among the vehicular nodes. Simulation results show that the proposed approach can increase the storage space utilization and has low data response time for vehicular networks. © 2017 Association for computing Machinery.

关键词： Digital storage

Differential Private Preservation Multi-core DBScan Clustering for Network User Data

学校读者我要写书评

暂无评论

Procedia Computer Science 2018年 129卷 257-262页

作者： Lina Ni Chao Li Haoran Liu Anu G. Bourgeois Jiguo Yu College of Computer Science and Engineering Shandong University of Science and Technology Qingdao China The Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai China Department of Computer Science Georgia State University University Plaza Atlanta GA 30303 USA School of Information Science and Engineering Qufu Normal University Shandong China

Mining rich network user data to extract valuable information becomes the focus and commercial interest of researchers. Differential privacy protection is a new paradigm based on data distortion, which protects sensitive data while maintains a certain statistical properties by adding random noise, and makes up for the shortcomings that traditional schemas require knowledge background assumptions and cannot analyze quantitatively. In this paper, we propose a DP-MCDBScan schema based on the powerful differential privacy. We perform simulation to evaluate our schema, whose results show that our schema has better efficiency, accuracy, and privacy protection effect than previous schemas.

关键词： Private preserving differential privacy MCDBScan clustering data mining Laplace noise

Full-Duplex Millimeter-Wave Communication

学校读者我要写书评

暂无评论

arXiv 2017年

作者： Xiao, Zhenyu Xia, Pengfei Xia, Xiang-Gen Beihang University Key Laboratory of Embedded System and Service Computing Tongji University University of Delaware

The potential of doubling the spectrum efficiency of full-duplex (FD) transmission motivates us to investigate FD millimeter-wave (FD-mmWave) communication. To realize FD transmission in the mmWave band, we first introduce possible antenna configurations for FD-mmWave transmission. It is shown that, different from the cases in micro-wave band FD communications, the configuration with separate Tx/Rx antenna arrays appears more flexible in self-interference (SI) suppression while it may increase some cost and area versus that with the same array. We then model the mmWave SI channel with separate Tx/Rx arrays, where a near-field propagation model is adopted for the line-of-sight (LOS) path, and it is found that the established LOS-SI channel with separate Tx/Rx arrays also shows spatial sparsity. Based on the SI channel, we further explore approaches to mitigate SI by signal processing, and we focus on a new cancellation approach in FD-mmWave communication, i.e., beamforming cancellation. Centered on the constant-amplitude (CA) constraint of the beamforming vectors, we propose several candidate solutions. Lastly, we consider an FD-mmWave multi-user scenario, and show that even if there are no FD users in an FD-mmWave cellular system, the FD benefit can still be exploited in the FD base station. Candidate solutions are also discussed to mitigate both SI and multi-user interference (MUI) simultaneously. Copyright © 2017, The Authors. All rights reserved.

关键词： Beamforming

Automatic vector-based road structure mapping using multi-beam lidar

学校读者我要写书评

暂无评论

arXiv 2018年

作者： He, Xudong Zhao, Junqiao Sun, Lu Huang, Yewei Zhang, Xinglian Li, Jun Ye, Chen Key Laboratory of Embedded System and Service Computing Tongji University Ministry of Education Shanghai Department of Computer Science and Technology School of Electronics and Information Engineering Tongji University Shanghai School of Surveying and Geo-Informatics Tongji University Shanghai China School of Automotive Studies Tongji University Shanghai China

In this paper, we studied a SLAM method for vector-based road structure mapping using multi-beam LiDAR. We propose to use the polyline as the primary mapping element instead of grid cell or point cloud, because the vector-based representation is precise and lightweight, and it can directly generate vector-based High-Definition (HD) driving map as demanded by autonomous driving systems. We explored: 1) the extraction and vectorization of road structures based on local probabilistic fusion. 2) the efficient vector-based matching between frames of road structures. 3) the loop closure and optimization based on the pose-graph. In this study, we took a specific road structure, the road boundary, as an example. We applied the proposed matching method in three different scenes and achieved the average absolute matching error of 0.07 m. We further applied the mapping system to the urban road with the length of 860 meters and achieved an average global accuracy of 0.466 m without the help of high precision GPS. Copyright © 2018, The Authors. All rights reserved.

关键词： Mapping

TiEV: The tongji intelligent electric vehicle in the intelligent vehicle future challenge of China

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Zhao, Junqiao Ye, Chen Wu, Yan Guan, Linting Cai, Lewen Sun, Lu Yang, Tao He, Xudong Li, Jun Ding, Yongchao Zhang, Xinglian Wang, Xinchen Huang, Jinlin Zhang, Enwei Huang, Yewei Jiang, Wei Zhang, Shaoming Xiong, Lu Feng, Tiantian Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai Department of Computer Science and Technology School of electrics and Information Engineering Tongji University Shanghai School of Surveying and Geo-Informatics Tongji University Shanghai China School of Automotive Studies Tongji University Shanghai China

TiEV is an autonomous driving platform implemented by the Tongji University of China. The vehicle is drive-by-wire and is fully powered by electricity. We devised the software system of TiEV from scratch, which is capable of driving the vehicle autonomously in urban paths as well as on fast express roads. We describe our whole system, especially novel modules of probabilistic perception fusion, large-scale mapping and updating, the 1st and the 2nd planning and the overall safety concern. TiEV finished 2016, and 2017 Intelligent Vehicle Future Challenge of China held at Changshu. We show our experiences on the development of autonomous vehicles and discuss the future trends. Copyright © 2018, The Authors. All rights reserved.

关键词： Autonomous vehicles

General-to-Specialized Analysis Based on Deep Belief Network

学校读者我要写书评

暂无评论

General-to-Specialized Analysis Based on Deep Belief Network

Asian Conference on Pattern Recognition (ACPR)

作者： Renjie Hu Lianghua He Yuqin Wang D. Hu Key Laboratory of Embedded System and Service Computing Tongji University Shanghai China Key Laboratory of EMW Information Fudan University Shanghai CN

Recently, the deep neural network has achieved great performance in many areas. During analysis, all learned features are used at once, some of which could bring negative affect to specific classes. Recently, cognitive studies show that a human visual cognition process is hierarchical and dynamic, i.e., when meeting different targets, human brain intends to pay attention to different parts. Therefore, in this paper, we introduce this kind of mechanism into deep belief network (DBN) and propose a new general-to-specialized algorithm. Firstly, hierarchical knowledge networks are constructed based on the original learned DBN through pruning and retraining. Because these networks are learned for different discriminate ability, we call them as the general network and the specialized network separately. Secondly, a general-to-specialized analysis method is proposed which is proved theoretically. When predicting the class of an input sample, we select the corresponding specialized networks according to the preliminary analysis result and then make in-depth analysis. Experiments on four benchmark datasets are performed to test the proposed algorithm. The results show that our algorithm is feasible, valid and robust.

关键词： Training Visualization Feature extraction Prediction algorithms Biological neural networks Cognition Recurrent neural networks