CoGCN: Combining co‐attention with graph convolutional network for entity linking with knowledge graphs

Entity linking is a fundamental task in natural language processing. The task of entity linking with knowledge graphs aims at linking mentions in text to their correct entities in a knowledge graph like DBpedia or YAGO2. Most of existing methods rely on hand‐designed features to model the contexts of mentions and entities, which are sparse and hard to calibrate. In this paper, we present a neural model that first combines co‐attention mechanism with graph convolutional network for entity linking with knowledge graphs, which extracts features of mentions and entities from their contexts automatically. Specifically, given the context of a mention and one of its candidate entities' context, we introduce the co‐attention mechanism to learn the relatedness between the mention context and the candidate entity context, and build the mention representation in consideration of such relatedness. Moreover, we propose a context‐aware graph convolutional network for entity representation, which takes both the graph structure of the candidate entity and its relatedness with the mention context into consideration. Experimental results show that our model consistently outperforms the baseline methods on five widely used datasets.     

A Network Security Risk Assessment Method Based on a B_NAG Model

Computer networks face a variety of cyberattacks. Most network attacks are contagious and destructive, and these types of attacks can be harmful to society and computer network security. Security evaluation is an effective method to solve network security problems. For accurate assessment of the vulnerabilities of computer networks, this paper proposes a network security risk assessment method based on a Bayesian network attack graph (B_NAG) model. First, a new resource attack graph (RAG) and the algorithm E-Loop, which is applied to eliminate loops in the B_NAG, are proposed. Second, to distinguish the confusing relationships between nodes of the attack graph in the conversion process, a related algorithm is proposed to generate the B_NAG model. Finally, to analyze the reachability of paths in B_NAG, the measuring indexs such as node attack complexity and node state transition are defined, and an iterative algorithm for obtaining the probability of reaching the target node is presented. On this basis, the posterior probability of related nodes can be calculated. A simulation environment is set up to evaluate the effectiveness of the B_NAG model. The experimental results indicate that the B_NAG model is realistic and effective in evaluating vulnerabilities of computer networks and can accurately highlight the degree of vulnerability in a chaotic relationship.     

Towards sustainable industry 4.0: A green real-time IIoT multitask scheduling architecture for distributed 3D printing services

As the keystones of the personalized manufacturing, the Industrial Internet of Things (IIoT) consolidated with 3D printing pave the path for the era of Industry 4.0 and smart manufacturing. By resembling the age of craft manufacturing, Industry 4.0 expedites the alteration from mass production to mass customization. When distributed 3D printers (3DPs) are shared and collaborated in the IIoT, a promising dynamic, globalized, economical, and time-effective manufacturing environment for customized products will appear. However, the optimum allocation and scheduling of the personalized 3D printing tasks (3DPTs) in the IIoT in a manner that respects the customized attributes submitted for each model while satisfying not only the real-time requirements but also the workload balancing between the distributed 3DPs is an inevitable research challenge that needs further in-depth investigations. Therefore, to address this issue, this paper proposes a real-time green-aware multi-task scheduling architecture for personalized 3DPTs in the IIoT. The proposed architecture is divided into two interconnected folds, namely, allocation and scheduling. A robust online allocation algorithm is proposed to generate the optimal allocation for the 3DPTs. This allocation algorithm takes into consideration meeting precisely the customized user-defined attributes for each submitted 3DPT in the IIoT as well as balancing the workload between the distributed 3DPs simultaneously with improving their energy efficiency. Moreover, meeting the predefined deadline for each submitted 3DPT is among the main objectives of the proposed architecture. Consequently, an adaptive real-time multi-task priority-based scheduling (ARMPS) algorithm has been developed. The built ARMPS algorithm respects both the dynamicity and the real-time requirements of the submitted 3DPTs. A set of performance evaluation tests has been performed to thoroughly investigate the robustness of the proposed algorithm. Simulation results proved the robustness and scalability of the proposed architecture that surpasses its counterpart state-of-the-art architectures, especially in high-load environments.     

基于时序逆影响的随机游走推荐算法*

相似性计算是协同过滤推荐的关键步骤,针对传统相似性计算认为相似关系是对等的且没有考虑消费顺序和时间间隔的问题,提出了基于时序逆影响的随机游走推荐算法。首先,基于用户时序关联图提出一种新的称为时序逆影响的相似性度量,利用随机游走得到了目标用户近邻集合;其次,利用随机游走在项目时序关联图上进一步改进推荐的多样性和覆盖率。它不但认为用户间相似是不对称的,考虑了用户消费项目的顺序和时间间隔,获得了用户全局的直接和间接近邻,而且考虑了项目间的时序逆影响。通过在真实数据集上的大量试验结果表明,与其他随机游走方法相比,不但能提高推荐性能、缓解数据稀疏,而且通过提高多样性和覆盖率解决了过拟合的问题。     

领域知识图谱快速构建和应用框架

快速构建和应用领域知识图谱(domain knowledge graphs, DKG)已成为企业的迫切需求。然而,领域知识图谱的快速构建和应用仍然存在问题:1）构建前,复杂领域场景涉及知识维度广,业务专家短时间内难以构建完备的领域schema;2）构建中,业务和技术专家深度耦合,图谱构建缓慢,难以适应业务快速发展需求;3）构建后,图谱应用严重依赖技术人员开发,业务专家无法自主基于领域知识图谱探索问题解决方案。为解决上述问题,本文提出了领域知识图谱的快速构建和应用框架,其中包括:多人协作构建领域schema解决领域知识的复杂性问题,将业务和技术专家解耦合,提高领域知识图谱的构建效率,最后通过建立基于行业schema的auto-KBQA((knowledge base question answering)解决领域知识图谱在知识问答应用的快速落地。通过实际项目的应用落地,验证了该框架可有效加快领域知识图谱的落地应用,期望该框架给领域知识图谱的快速构建和应用带来一定的启发和帮助。     

基于图嵌入的自适应多视降维方法

随着监控摄像头的普及和数据采集技术的快速发展,多视数据呈现出规模大、维度高和多源异构的特点,使得数据存储空间大、传输慢、算法复杂度高,造成“有数据、难利用”的困境。到目前为止,国内外在多视降维方面的研究还比较少。针对这一问题,本文提出一种基于图嵌入的自适应多视降维方法。该方法在考虑视角内降维后数据重构原始高维数据的基础上,提出自适应学习相似矩阵来探索不同视角之间降维后数据的关联关系,学习各视数据的正交投影矩阵实现多视降维任务。本文在多个数据集上对降维后的多视数据进行了聚类/识别实验验证,实验结果表明基于图嵌入的自适应多视降维方法优于其他降维方法。     

基于图结构源代码切片的智能化漏洞检测系统

针对智能化漏洞检测,从源代码程序依赖图中根据漏洞特征提取图结构源代码切片,将图结构切片信息表征后利用图神经网络模型进行漏洞检测工作。实现了切片级的漏洞检测,并在代码行级预测漏洞行位置。为了验证系统的有效性,分别与静态漏洞检测系统、基于序列化文本信息和基于图结构化信息的漏洞检测系统做比较,实验结果表明,所提系统在漏洞检测能力上有较高准确性,并且在漏洞代码行预测工作上有较好表现。     

基于结构熵约束的图聚类方法

针对以大数据为中心的信息开放共享平台,如何从嵌入大规模噪声结构的网络中解码出网络的真实结构,进一步在挖掘关联信息的过程中得到较为准确的挖掘结果的问题,提出基于结构熵的聚类方法实现对图中节点关联程度的划分。提出了计算二维结构信息的求解算法和基于熵减原则的模块划分算法,对图结构中节点划分得到对应的模块;利用 K 维结构信息算法对已划分的模块做进一步的划分,实现对图结构中节点的聚类;通过实例分析表明,所提出的图聚类方法不仅能够反映图结构的真实结构,而且可以有效地挖掘出图结构中节点之间的关联程度。同时对比了其他3种聚类方法,实验表明该方法在执行时间上具有更高的效率和保证聚类结果的可靠性。     

基于图表示学习的消息回复关系判断方法

微信、QQ和钉钉等社交媒体都提供多对多聊天群组功能,这些聊天群组包含海量信息,对群组聊天内容进行有效分析,获取有价值的关联信息,是当前领域的研究热点。群组中用户间交互是群组实现的主要功能,用户间消息回复是用户间交互实现的方式,消息间的回复行为下隐藏着消息间和用户间的关系。群组消息间回复通常是隐式和非连续的,大部分群组消息间没有指定明确的回复关系,当前消息也不一定是上一条临近消息的回复,回复关系要根据具体的聊天场景确定。当消息间没有显示指定回复关系时,回复不易于分析和理解群组聊天内容,阻碍了对群组聊天内容的整体性分析。本论文针对群组消息间的回复关系,提出了基于图表示学习的消息回复关系判断方法,该方法不同于以往方法仅使用部分群组要素,是在综合学习消息的文本信息、发送消息的用户信息和上下文信息的基础上,根据群组内容构建群组图和生成自适应消息图,得到了多种群组要素信息和要素间关系组成的图结构,利用图模型在图结构上进行群组消息的表示学习,图模型输出群组消息的表示向量,拼接消息对的表示向量并进一步预测群组消息间的回复关系。在消息间回复关系的学习过程中,图模型通过任务学习更新图中消息节点,同时更新图中用户节点向量表示,经过用户向量分析实验验证了该模型输出的用户向量的有效性和合理性。在公开数据集和标注数据集上进行了对比实验和显著性检验分析,结果显示模型在多个评估指标上大幅优于对比模型,如在F1指标上,比单纯依赖BERT的句子对分类模型提高了接近20%。     

On the Girth of Tanner (5,7) Quasi-Cyclic LDPC Codes

Tanner (5,7)准循环LDPC码的围长问题

徐恒舟,白宝明,冯丹,孙成

(西安电子科技大学 ISN国家重点实验室,西安 710071)

创新点说明：

1）定义了Tanner (5,7)准循环LDPC码Tanner图中环的等价类;

2）提出了Tanner (5,7)准循环LDPC码Tanner图中环存在的充要条件,即素域Fp上多项式是否有一个35次单位根;

3）完全解决了Tanner (5,7)准循环LDPC码的围长问题（码长为7p,p为模35余1的素数）

研究目的：解决Tanner (5,7)准循环LDPC码的围长问题（码长为7p,p为模35余1的素数）

研究方法：

本文定义了环的等价形式,并分析了Tanner (5,7)准循环LDPC码Tanner图中的环结构,并将长度为4,6,8和10的环划分为16个等价类。此外,这些环是否存在可以等价地看为素域Fp上多项式是否有一个35次单位根。通过检验这些多项式是否存在一个35次单位根,可以得到Tanner (5,7)准循环LDPC码的围长候选值。最后,统计这些候选值得到了Tanner (5,7)准循环LDPC码的围长。

结 果：

得到了围长为6,8和10的Tanner (5,7)准循环LDPC码,其中码长为7p：

1）当围长为6时,p = 71。

2）当围长为8时,p ? G8 = {211, 281, 421, 491, 631, 701, 911, 1051, 2311, 4271, 5531, 7211, 237301, 354551}。

3） 当围长为10时,p ? G10 = {1471, 2381, 2521, 2591, 2731, 2801, 3011, 3221, 3361, 3571, 3851, 4201, 4481, 4621, 4691, 4831, 5741, 5881, 6091, 6301, 6581, 6791, 7001, 7351, 7561, 7841, 8191, 8681, 8821, 9241, 9311, 9521, 9661, 9871, 9941, 10151, 10501, 10711, 10781, 11131, 11411, 11621, 11831, 11971, 12041, 12251, 12391, 12601, 12671, 13441, 13931, 14071, 14771, 15121, 15541, 16381, 16451, 16661, 16871, 17011, 17291, 17431, 17921, 18061, 18131, 18481, 18691, 19181, 19391, 19531, 20161, 20231, 20441, 21001, 21211, 21491, 21701, 21911, 22051, 22751, 24151, 24781, 25411, 26111, 26251, 28001, 28771, 30661, 30871, 30941, 32971, 33181, 33461, 33811, 34231, 34511, 35141, 36541, 37871, 38011, 39551, 39761, 42491, 43261, 43331, 44171, 45361, 46831, 47041, 47741, 47881, 48371, 50051, 51521, 52361, 54881, 55511, 55721, 57751, 59221, 63841, 65101, 66571, 66851, 67061, 67271, 71191, 74761, 75181, 76231, 79801, 85751, 97441, 98491, 104021, 109831, 110321, 110951, 112771, 118861, 122921, 125231, 126211, 127261, 128591, 130621, 134401, 137131, 141961, 147211, 152041, 154981, 159671, 162821, 164431, 185221, 192431, 203911, 204331, 207061, 217351, 242621, 262781, 273001, 274471, 278741, 280351, 285251, 296731, 299671, 301841, 318641, 325921, 333691, 343141, 343561, 348461, 349931, 361901, 370441, 374291, 385631, 393961, 403621, 423431, 435401, 437501, 440651, 441421, 443591, 446881, 453461, 495461, 522061, 532421, 557831, 589471, 687541, 704761, 718271, 763771, 766501, 829151, 837271, 845951, 867371, 898661, 920641, 1022141, 1180901, 1197281, 1239421, 1253071, 1388381, 1542031, 1634011, 1747271, 1773241, 2102171, 2153551, 2318471, 2691011, 3338441, 3439801, 4567151, 4649261, 8553581, 9268631, 23632351, 27136621}。

结 论：

当p ? P35/({71}? G8?G10)时,Tanner (5,7)准循环LDPC码的围长为12（P35代表模35余1的素数集合）。