工业互联网正在加速推进,其中数据隐私泄露、数据安全和数据追溯等问题严重制约了企业之间的信息流通和数据共享.区块链作为一种去中心化、无需中介、不可篡改、可追溯的技术,在金融、虚拟货币领域被广泛应用.应用区块链技术可以很好地解决工业互联网的应用难题.本文详细介绍工业区块链的内涵和应用现状,重点分析了工业区块链的核心技术和工业应用面临的难题,并对国内外在相关技术方面的研究进展进行了介绍.

The rapid advances in positioning technology have created huge spatio-temporal trajectory data, and there are always obvious aberrant outliers in trajectory data. Detecting outliers in the trajectory is critical to improving data quality and the accuracy of subsequent trajectory data mining tasks. In this paper, we propose a trajectory outlier detection algorithm based on a Bidirectional Long Short-Term Memory (Bi-LSTM) model. First, a six-dimensional motion feature vector is extracted for each trajectory point, and then we construct a Bi-LSTM model. The model input is the trajectory data feature vector of a certain sequence length, and its output is the class type of the current track point. In addition, a combination method of undersampling and oversampling is applied to mitigate the effect of data distribution imbalance on detection performance. The Bi-LSTM model can automatically learn the difference between the normal points and adjacent abnormal points in the motion characteristics by combining the LSTM unit and the bidirectional network. Experimental results based on a real ship trajectory annotation data show that the detection performance of our proposed algorithm significantly exceeds those of the constant velocity threshold algorithm, non-sequential classical machine learning classification algorithms, and convolutional neural network model. Especially, the recall value of the proposed algorithm reaches 0.902, which verifies its effectiveness.

Traditional supervised learning algorithms of instance alignment depend on large amounts of labeled data,and the feature representation methods are not suitable for data in encyclopedia.In view of these issues,a semi-supervised co-training instance alignment method is proposed.Instance alignment is modeled as a constrained binary classification problem.Then multiple features are extracted by fully utilizing different categories of existing information,including instance names,attributes,description texts and the critical discrete values extracted from the texts,such as temporal and numerical values.The features are divided into two relatively independent views,and two models are trained interactively on these two views,in order to learn more about the distribution of synonymous instances from the unlabeled data iteratively.Experimental results between two Chinese encyclopedia datasets show that the proposed method achieves a 84.3％ F1-value on aligning instances,and outperforms other comparative methods,proving the effectiveness and applicability of the semi-supervised co-training instance alignment method.

The complex network is becoming increasingly concerned by the experts and scholars.The visualization of the complex network can help the users to discover the hidden knowledge and information in the complex system represented by complex network,which is of great significance to the fields of computer science,sociology,and biology.The force-directed layout algorithm is the mainstream algorithm in the field of complex network visualization.It uses the form of node connection graph to abstract the complex network,the layout follows aesthetic standards such as the uniform distribution of nodes and the uniform of edges,to a certain extent,which hinders the display of the community structure of complex networks.Aiming at above problems,this paper introduces the repulsion and gravitational force of the community based on the degree centrality to improve the clustering layout of the complex network.The experimental results show that the proposed algorithm can effectively display the community structure of complex networks while preserving the information of margin nodes between communities.

近年来,基于知识图谱的问答系统逐渐成为学术界和工业界的研究和应用热点方向,而传统方法通常存在效率不高以及未充分利用数据信息的问题。针对以上问题,本文将中文知识图谱问答分为实体抽取和属性选择2个子任务,采用双向长短期记忆条件随机场(Bi-LSTM-CRF)模型来进行实体识别,并提出一种多粒度特征表示的属性选择模型。该模型采用字符级别以及词级别分别对问句和属性进行嵌入表示并通过编码器进行编码,对于属性同时还引入热度编码的信息。通过不同粒度文本表示的结合,并对问句和属性进行相似度计算,最终该系统在NLPCC-ICCPOL 2016 KBQA数据集上取得了73.96%的F1值,能够较好地完成知识图谱问答任务。

定位技术的快速发展催生了轨迹大数据,轨迹数据中总是存在着明显偏离轨迹的异常点。检测出轨迹中的异常点对提高数据质量和后续知识发现精度至关重要。目前轨迹异常点检测算法主要为恒定速度阈值法,没有考虑目标在不同时刻运动状态的变化,仅能检测出速度超出指定阈值的一部分异常点,甚至出现检测错误的情况,算法鲁棒性较差。针对现有问题,本文提出一种基于自适应阈值的轨迹异常点检测算法(Trajectory Outlier Detection Algorithm based on adaptive Threshold,TODAT)。TODAT算法充分考虑了目标在一段时间内的运动信息和观测噪声的影响,采用局部阈值窗和均值滤波窗来计算阈值和速度,同时又引入了经济航速阈值和连续异常点放回机制。基于真实船舶数据的实验结果表明,本文算法可根据轨迹数据得到自适应的阈值,有效检测出全部异常点,大幅度提高轨迹数据的质量。

The current time-varying network layout algorithm mainly preserves the mental graph from the network structure and the aesthetic quality,without considering the influence of the node centrality.The forcedirected algorithm to the time-varying network layout algorithm by embeding node centrality is improved.Firstly,the PageRank algorithm is used to evaluate the centrality of the nodes by referring to semi-local centrality.Then,the dynamic range is calculated as the constraint factor of nodes according to its centrality and stability.Finally,the time-varying network layout algorithm based on node centrality is formed by the static force-directed algorithm via the constraint of the dynamic range.The experimental results show that the proposed approach can balance the mental map and aesthetic quality,and also own good computational capability for large time-varying networks.

In order to improve the performance of property alignment between heterogeneous Chinese open datasets, a data-driven method for fine-grained alignment is proposed, which exploits the extension and domain information of properties to find equivalence, subsumption and relevance relations between properties in a unified way.First, the data types of properties are determined utilizing statistical theory, and a type-aware metric is given to calculate the similarity of properties.Based on that, the property relation recognition is modeled as a multi-classification problem, and effective features are generated to represent different property relationships and construct the random forest classifier.The experimental results show that, the proposed method can reach a precision of 94.6% in determining data types of properties, and the final F1 measures in recognizing equivalent, subsumptive and relevant properties are 71.3%, 57.3% and 59.9%, respectively.Compared with the traditional approaches that only focus on equivalent properties, the fine-grained property alignment method can improve the precision in recognizing equivalent properties, and recognize subsumptive and relevant properties, proving its effectiveness on Chinese open datasets.

In this paper, we study the characteristics and components of mobile wireless mesh backbone networks (MWMBNs). How to maintain the heterogeneous network connectivity in mobile scenarios is important. We propose a topology control algorithm based on the group mobility model for MWMBN to maintain network connectivity in mobile scenarios. It includes the following work: MWMBNs is classified into the mobile wireless backbone node group (MWBNG) and unit groups (UGs). the MWBNG is composed of all the MWBNs while the UG is composed of a MWBN and several TNs connect it. In the MWBNG, we propose the reference points assignment algorithm (RPAA) based on the RPGM model. In the UG, the TNs has different communication radius with the MWBNs', we propose the topology control algorithm based on tolerate zone using nomadic community mobility, the control algorithm does not need to use any control information to maintain topology control or to change the routing table. The TNs change the self-position to maintain the connectivity with the MWBN. The simulation results show the topology control algorithm based on the group mobility model for MWMBN to maintain the heterogeneous network connectivity when they move together.

针对k均值算法在文本聚类中由于初始聚类质心随机选择,使得聚类结果陷入局部最优,且孤立点和不确定的聚类个数造成k均值算法准确性低、收敛速度慢的问题,提出了一种改进的k均值文本聚类算法.该算法采用fp-growth算法挖掘文本频繁项集,过滤频繁项集得到核心频繁项集,并利用核心频繁项集指导文本初始聚类质心和聚类个数的生成,最后k均值算法利用初始聚类质心和聚类个数完成文本聚类.在新浪微博数据集上进行文本聚类实验,实验结果表明,改进的k均值算法提高了文本聚类的准确性,加快了收敛速度,具有较强的鲁棒性.

In response to the problem of how to guide the router nodes to move, a novel router node deci-sion-making mechanism was proposed. This mechanism includes the classification of mobile router nodes, the state transition of router nodes, the movement decision of router nodes, and the topology optimization. The simulation results show that the wireless mobile Mesh network using this mechanism can prevent the network partitioning and provide higher forwarding throughput for the mobile client nodes.

针对二分网络中社区挖掘的准确性不高、对额外参数的依赖较大的问题,基于谱聚类算法的思想,从二分网络的拓扑结构展开,提出了一种改进的社区挖掘算法。该算法将二分网络映射到单一网络进行社区挖掘,采用资源分布矩阵替代传统的邻接矩阵,挖掘出同类节点间的隐含信息,有效地保证了原图的信息,改进了谱聚类算法的输入,提高了社区挖掘的准确性;将模块度函数概念应用到聚类分析中,用模块度衡量社区挖掘的质量,有效解决了自动确定聚类数目的问题。在实际网络和人造网络上进行实验,与蚁群优化算法、边集聚系数算法等算法进行对比,实验结果表明,所提算法不但能较准确地获得二分网络的社区数目,且在不需要任何额外参数的情况下,能获得很好的划分效果,可以应用于深入理解二分网络,进行推荐、影响力分析等。

In order to achieve secure storage and transmission of digital images,an image encryption technology based on the genetic algorithm is proposed.Secret key sequence generated is more random and complex by this technique.An image can be transformed into another image,so as to achieve the original image encryption.The Matlab experimental results show that the algorithm is effective and feasible.

Due to most of the existing approaches for event extraction generally cause an imbalance between positive and nega-tive samples,a new extraction method driven by event instance is presented.The method removes the non-event sentences with support vector machine,and then a novel distance metric is presented which uses Markov model to describe the location of the sentences in documents.Experimental results indicate that the imbalance problem can be solved and the overall per-formance of event extraction is improved.

Most of the traditional approaches for automatic image annotation generally demand on training data with exact labels .However ,this kind of data usually comes from human generated tags ,which are too subjective and often difficult to obtain .In this paper ,we propose a novel image annotation model which can utilize the rich surrounding text of images by integrating entity extraction technology of Natural Language Processing .Then the image features ,surrounding text and extracted entity words which we assume that can more directly describe the salient objects in the corresponding ima‐ges are modeled in a uniform probabilistic topic model framework .The learned correlations among different data modali‐ties can be used in image annotation afterwards .Experimental results on UIUC Pascal Sentence dataset show that our model outperforms the traditional topic model‐based image annotation methods in both annotation and retrieval .

To accurately analyze and measure the topical influence of micro-blog users,a new influence evaluation algorithm based on multi-relationship and personal attributes is presented.Taking topic as unit,the multi-relational influence network including repost,comment,copy and mention relations is constructed by analyzing user interaction behavior,and the transi-tion probability calculation model is given.In consideration of user activity,user authority,micro-blog quality and follower quality,an index system of user’s personal attributes influence based on the topic is constructed,and the weight of each in-dicator is given by using the method of analytic hierarchy process (AHP).Finally,a calculation method of influence based on PageRank is designed by fusing user relations and attributes.A large number of experiments according to real data sets show that the proposed method is accurate and effective.

Most of the traditional approaches for image annotation based on distance metric learning generally suppose that the constraints on training data is explicit,which only works on small datasets with exact labels.As the scale of the dataset becomes larger and most images are accompanied by noisy labels,this ideal assumption will be not efficient.In this paper,we propose a novel distance metric learning method based on probabilistic topic model.The uncertain and latent side information can be mined by a probabilistic topic model and afterwards used in distance metric learning.The learned semantic distance metric then can be used in the search-based image annotation.Experiments on Flickr dataset demonstrate that the proposed model outperforms the state-of-the-art annotation methods.

Due to the short message,scattering format and content,incomplete description and data noise,the micro-blog topic can not be obtained efficiently.An improved SMLDA model based on LDA is proposed.The model mainly takes mi-cro-blog authors association,the specific topic tags,the relay document association and the background topic into considera-tion and adopts the Gibbs sampling algorithm to derive parameters.The experimental results on Sina micro-blog data set show that compared with LDA model,the SMLDA model is more effective.

针对无线传感器网络(WSNs)中无线信号动态波动变化时引起网络链路不稳定和覆盖范围变小的问题,提出了一种基于TopDisc的WSNs拓扑控制算法.该算法通过引入拓扑控制参数控制网络拓扑优化以适应无线信号变化,给出了拓扑创建过程,对算法进行了仿真实验.仿真结果表明:改进的拓扑控制算法能够提升WSNs对复杂无线电环境的适应能力,提高了资源利用率和链路可靠性.

ECG data acquisition instruments are bulky,high power,low mobility and endurance,so a kind of portable ECG acquisition system based on MSP430 is designed.With the main control chip MSP430F149,the system gets two 24 bits of data from special bioelectricity sampling chip ADS1291 front end through the SPI interface.After decoding,filtering and re-coding,data is stored in the external flash chip,at the same time it will be processed via a serial port and sent to HC-05 Bluetooth module,and extended to mobile terminals with Bluetooth communication functions.The test results show that the system can accurately measure the first lead electrocardiogram (ECG)of human body,has small volume,low power con-sumption,high accuracy and stability,etc.