异质信息网络高阶层次化嵌入学习与推荐预测

异质信息网络是一种异质数据表示形式,如何融合异质数据复杂语义信息,是推荐系统面临的挑战之一.利用弱关系具有的丰富语义和信息传递能力,构建一种面向推荐系统的异质信息网络高阶嵌入学习框架,主要包括:初始化信息嵌入、高阶信息嵌入聚合与推荐预测3个模块.初始化信息嵌入模块首先采用基于弱关系的异质信息网络最佳信任路径筛选算法,有效地避免在全关系异质信息网络中,采样固定数量邻居造成的信息损失,其次利用新定义的基于多头图注意力的多任务共享特征重要性度量因子,筛选出节点的语义信息,并结合交互结构,有效地表征网络节点;高阶信息嵌入聚合模块通过融入弱关系及网络嵌入对知识良好的表征能力,实现高阶信息表达,并利用异质信息网络的层级传播机制,将被采样节点的特征聚合到待预测节点;推荐预测模块利用高阶信息的影响力推荐方法,实现了推荐任务.该框架具有嵌入节点类型丰富、融合共享属性和隐式交互信息等特点.最后,实验验证UI-HEHo学习框架可有效地改善评级预测的准确性,以及推荐生成的针对性、新颖性和多样性,尤其是在数据稀疏的应用场景中,具有良好的推荐效果.     

Rumor detection with self-supervised learning on texts and social graph

Rumor detection has become an emerging and active research field in recent years. At the core is to model the rumor characteristics inherent in rich information, such as propagation patterns in social network and semantic patterns in post content, and differentiate them from the truth. However, existing works on rumor detection fall short in modeling heterogeneous information, either using one single information source only (e.g., social network, or post content) or ignoring the relations among multiple sources (e.g., fusing social and content features via simple concatenation).Therefore, they possibly have drawbacks in comprehensively understanding the rumors, and detecting them accurately. In this work, we explore contrastive self-supervised learning on heterogeneous information sources, so as to reveal their relations and characterize rumors better. Technically, we supplement the main supervised task of detection with an auxiliary self-supervised task, which enriches post representations via post self-discrimination.Specifically, given two heterogeneous views of a post (i.e., representations encoding social patterns and semantic patterns), the discrimination is done by maximizing the mutual information between different views of the same post compared to that of other posts. We devise cluster-wise and instance-wise approaches to generate the views and conduct the discrimination, considering different relations of information sources. We term this framework as self-supervised rumor detection (SRD). Extensive experiments on three real-world datasets validate the effectiveness of SRD for automatic rumor detection on social media.     

Measuring code maintainability with deep neural networks

The maintainability of source code is a key quality characteristic for software quality. Many approaches have been proposed to quantitatively measure code maintainability. Such approaches rely heavily on code metrics, e.g., the number of Lines of Code and McCabe’s Cyclomatic Complexity. The employed code metrics are essentially statistics regarding code elements, e.g., the numbers of tokens, lines, references, and branch statements. However, natural language in source code, especially identifiers, is rarely exploited by such approaches. As a result, replacing meaningful identifiers with nonsense tokens would not significantly influence their outputs, although the replacement should have significantly reduced code maintainability. To this end, in this paper, we propose a novel approach (called DeepM) to measure code maintainability by exploiting the lexical semantics of text in source code. DeepM leverages deep learning techniques (e.g., LSTM and attention mechanism) to exploit these lexical semantics in measuring code maintainability. Another key rationale of DeepM is that measuring code maintainability is complex and often far beyond the capabilities of statistics or simple heuristics. Consequently, DeepM leverages deep learning techniques to automatically select useful features from complex and lengthy inputs and to construct a complex mapping (rather than simple heuristics) from the input to the output (code maintainability index). DeepM is evaluated on a manually-assessed dataset. The evaluation results suggest that DeepM is accurate, and it generates the same rankings of code maintainability as those of experienced programmers on 87.5% of manually ranked pairs of Java classes.     

基于VoxelMorph的可见光和红外遥感影像配准

由于可见光和红外的成像机理、成像波段不同,获取的遥感影像之间存在复杂的非线性辐射畸变,传统的配准方法难以实现两者的高精度配准。本文提出一种基于VoxelMorph的可见光和红外遥感影像配准方法,利用卷积神经网络对可见光和红外异源图像进行分步的精细化形变场计算,从而实现快速高精度配准。将可见光图像作为参考图像,利用U-Net网络计算待配准红外图像和参考（可见光）图像的形变场,实现全局对齐的仿射变换,然后通过空间转换网络进一步实现更高自由度变形。采用WHU-OPT-SAR数据集的实验结果表明,与基于尺度不变特征变换（SIFT）算法的传统配准方法相比,本文提出的基于VoxelMorph配准方法可以获得更好的配准效果,验证了基于VoxelMorph的配准方法在多源遥感影像领域的有效性。     

基于信号特征提取和卷积神经网络的轴承故障诊断研究

轴承是机械设备主要零部件之一,也是机械设备主要故障零部件之一。轴承故障问题为机械设备的重点,机械设备的使用受到故障轴承的直接影响。针对传统的卷积神经网络算法轴承故障诊断效率低下问题,本文提出了一种基于信号特征提取和卷积神经网络的优化方法。首先对原始数据信号进行时域和频域的信号特征提取,获得有效的故障特征值。之后,使用卷积神经网络对提取的特征值进行故障诊断,完成故障分类。本文使用美国凯斯西储大学的滚动轴承振动加速度信号作为数据集,对提出的方法进行验证,得到的故障诊断平均准确率为74.37%,准确率的方差为0.0001;传统的卷积神经网络算法故障诊断平均准确率为65.6%;准确率的方差为0.0019。实验结果表明,相比传统的卷积神经网络,提出的方法对轴承故障诊断的准确率有显著的提高,并且该方法的稳定性更佳,计算时间更少,综合性能更佳。     

Volt-VAR control for distribution networks with high penetration of DGs: An overview

Continuously improving the ability to accept distributed renewable energies is the trend of future grid development, and a large number of papers have been published in recent years to study the problem of Volt-VAR control (VVC) for distribution networks with high penetration of distributed generations. This paper summarizes the relevant modeling and solution methods for VVC problems, mainly including VVC based on multiple time scales, hierarchical partitioning, multi-stage and network reconstruction, in conjunction with the operational characteristics of distribution networks containing distributed renewable energies; meanwhile, it analyzes the advantages and disadvantages of traditional optimization methods, heuristic intelligent algorithms and random variable processing methods used to solve VVC problems, and then introduces the application of model-free deep reinforcement learning as a latest decision method in VVC of distribution networks. Most of the models and methods compiled in this article are from the research results of the last three years and have some reference value.     

MC/DC Test Data Generation Algorithm Based on Whale Genetic Algorithm

The automatic generation of test data is a key step in realizing automated testing. Most automated testing tools for unit testing only provide test case execution drivers and cannot generate test data that meets coverage requirements. This paper presents an improved Whale Genetic Algorithm for generating test data required for unit testing MC/DC coverage. The proposed algorithm introduces an elite retention strategy to avoid the genetic algorithm from falling into iterative degradation. At the same time, the mutation threshold of the whale algorithm is introduced to balance the global exploration and local search capabilities of the genetic algorithm. The threshold is dynamically adjusted according to the diversity and evolution stage of current population, which positively guides the evolution of the population. Finally, an improved crossover strategy is proposed to accelerate the convergence of the algorithm. The improved whale genetic algorithm is compared with genetic algorithm, whale algorithm and particle swarm algorithm on two benchmark programs. The results show that the proposed algorithm is faster for test data generation than comparison methods and can provide better coverage with fewer evaluations, and has great advantages in generating test data.     

基于BP神经网络与改进粒子群的光伏MPPT算法

针对光伏阵列在环境突变情况下尤其是局部阴影下的多峰值现象,提出一种基于反向传播(BP)神经网络与改进粒子群的最大功率点跟踪(MPPT)算法。该算法利用BP神经网络近似定位最大功率点,并利用对粒子群算法中的惯性权重值进行非线性动态优化后的改进粒子群精确定位最大功率点。仿真结果表明,复合算法可以更好地跟踪最大功率点,有效避免前期易陷入局部极值的问题,提高了精度,减小了功率振荡。     

Clustering-based adaptive low-power subframe configuration with load-aware offsetting in dense heterogeneous networks

Heterogeneous Networks (HetNets) and cell densification represent promising solutions for the surging data traffic demand in wireless networks. In dense HetNets, user traffic is steered toward the Low-Power Node (LPN) when possible to enhance the user throughput and system capacity by increasing the area spectral efficiency. However, because of the transmit power differences in different tiers of HetNets and irregular service demand, a load imbalance typically exists among different serving nodes. To offload more traffic to LPNs and coordinate the Inter-Cell Interference (ICI), Third-Generation Partnership Project (3GPP) has facilitated the development of the Cell Range Expansion (CRE), enhanced Inter-Cell Interference Coordination (eICIC) and Further enhanced ICIC (FeICIC). In this paper, we develop a cell clustering-based load-aware offsetting and an adaptive Low-Power Subframe (LPS) approach. Our solution allows the separation of User Association (UA) functions at the User Equipment (UE) and network server such that users can make a simple cell-selection decision similar to that in the maximum Received Signal Strength (max-RSS) based UA scheme, where the network server computes the load-aware offsetting and required LPS periods based on the load conditions of the system. The proposed solution is evaluated using system-level simulations wherein the results correspond to performance changes in different service regions. Results show that our method effectively solves the offloading and interference coordination problems in dense HetNets.     

基于自注意力对抗的深度子空间聚类

子空间聚类(Subspace clustering)是一种当前较为流行的基于谱聚类的高维数据聚类框架.近年来,由于深度神经网络能够有效地挖掘出数据深层特征,其研究倍受各国学者的关注.深度子空间聚类旨在通过深度网络学习原始数据的低维特征表示,计算出数据集的相似度矩阵,然后利用谱聚类获得数据的最终聚类结果.然而,现实数据存在维度过高、数据结构复杂等问题,如何获得更鲁棒的数据表示,改善聚类性能,仍是一个挑战.因此,本文提出基于自注意力对抗的深度子空间聚类算法(SAADSC).利用自注意力对抗网络在自动编码器的特征学习中施加一个先验分布约束,引导所学习的特征表示更具有鲁棒性,从而提高聚类精度.通过在多个数据集上的实验,结果表明本文算法在精确率(ACC)、标准互信息(NMI)等指标上都优于目前最好的方法.