Multi-label dimensionality reduction and classification with extreme learning machines

In the need of some real applications, such as text categorization and image classification, the multi-label learning gradually becomes a hot research point in recent years. Much attention has been paid to the research of multi-label classification algorithms. Considering the fact that the high dimensionality of the multi-label datasets may cause the curse of dimensionality and wil hamper the classification process, a dimensionality reduction algorithm, named multi-label kernel discriminant analysis （MLKDA）, is proposed to reduce the dimensionality of multi-label datasets. MLKDA, with the kernel trick, processes the multi-label integrally and realizes the nonlinear dimensionality reduction with the idea similar with linear discriminant analysis （LDA）. In the classification process of multi-label data, the extreme learning machine （ELM） is an efficient algorithm in the premise of good accuracy. MLKDA, combined with ELM, shows a good performance in multi-label learning experiments with several datasets. The experiments on both static data and data stream show that MLKDA outperforms multi-label dimensionality reduction via dependence maximization （MDDM） and multi-label linear discriminant analysis （MLDA） in cases of balanced datasets and stronger correlation between tags, and ELM is also a good choice for multi-label classification.     

Optimized LDPC differential-encoded 16QAM scheme for high-speed transmission systems

The 16-ary quadrature amplitude modulation （16QAM） is a high spectral efficient scheme for high-speed transmission systems. To remove the phase ambiguity in the coherent detection system, differential-encoded 16QAM （DE-16QAM） is usually used, however, it will cause performance degradation about 3 dB as compared to the conventional 16QAM. To overcome the performance loss, a serial concatenated system with outer low density parity check （LDPC） codes and inner DE-16QAM is proposed. At the receiver, joint iterative differential demodulation and decoding （ID） is carried out to approach the maximum likelihood performance. Moreover, a genetic evolution algorithm based on the extrinsic information transfer chart is proposed to optimize the degree distribution of the outer LDPC codes. Both theoretical analyses and simulation results indicate that this algorithm not only compensates the performance loss, but also obtains a significant performance gain, which is up to 1 dB as compared to the conventional non-DE-16QAM.     

信息物理融合系统可信软件形式化建模与分析

从多Agent系统的角度,以Petri网和π演算为语义基础,建立了一种信息物理融合系统（cyber-physical systems,CPS）可信软件形式化模型（high-confidence software formal model,HCSFM）. HCSFM以Petri网形象地描述CPS可信软件静态结构模型及动态行为,用Petri网分析方法和支持工具对模型进行分析和验证; 利用π演算刻画CPS可信软件中Agent的加入、退出、更新和体系结构重配置等动态演化机制,并研究Agent的演化策略及演化后CPS的一致性,确保动态演化后CPS软件能正常交互,从而为CPS软件设计提供可信保障. 通过HCSFM在无人驾驶车辆编队CPS中的应用,表明HCSFM可以有效地对CPS可信软件进行建模和分析.     

The distributed representation for societal risk classification toward BBS posts

The risk classification of BBS posts is important to the evaluation of societal risk level within a period. Using the posts collected from Tianya forum as the data source, the authors adopted the societal risk indicators from socio psychology, and conduct document-level multiple societal risk classification of BBS posts. To effectively capture the semantics and word order of documents, a shallow neural network as Paragraph Vector is applied to realize the distributed vector representations of the posts in the vector space. Based on the document vectors, the authors apply one classification method KNN to identify the societal risk category of the posts. The experimental results reveal that paragraph vector in document-level societal risk classification achieves much faster training speed and at least 10% improvements of F-measures than Bag-of-Words. Furthermore, the performance of paragraph vector is also superior to edit distance and Lucene-based search method. The present work is the first attempt of combining document embedding method with socio psychology research results to public opinions area.     

脉冲测量雷达卫星标校的实现

针对轴系误差标定是脉冲测量雷达使用维护的重要内容,是确保其测量精度的主要手段。介绍了脉冲测量雷达轴系误差的卫星标定方法,给出了标校模型和一种基于最速下降法的迭代算法,结合某雷达的精度鉴定进行了应用研究。结果表明,卫星标定方法完全可行,可以避免常规标定方法所涉及的大量人工参与,标校结果更为客观、真实、可信,并分析了卫星标校方法的使用条件。     

基于压缩感知测量值的扩展目标检测

针对宽带雷达扩展目标检测这一问题,提出了一种在复高斯白噪声背景下基于压缩感知(compressed sensing, CS)测量值的检测新方法。新方法将CS理论引入到宽带雷达扩展目标检测领域,首先通过构造sinc基来稀疏表示扩展目标的一维距离像(high resolution range profile, HRRP),再由复杂的近似消息传递(complex approximate message passing, CAMP) 算法从被复高斯白噪声污染的CS测量值中得到HRRP由sinc基线性表示的相关系数,最后由基于l-0范数的检测器实现扩展目标检测,同时经过推导得到虚警概率和检测概率。基于实测宽带雷达回波数据的实验结果表明,所构造的sinc基可以较好地稀疏表示扩展目标的HRRP;和传统的检测器相比,所提出的新方法可以更好地实现扩展目标检测。     

自主空中加油输油阶段无人机建模与控制

针对自主空中加油输油阶段无人机位置保持控制问题,将无人机分为飞机本体和油箱两部分,利用变质量系统理论建立了无人机非线性方程,解决了常规模型无法反映出无人机重量、重心变化的问题。控制律设计方面,通过将变化的油箱转化为外界干扰,提出了基于干扰观测器控制（disturbance observer based control, DOBC）的复合控制结构。复合控制器由位置保持控制器和干扰观测器组成,位置保持控制器采用带积分的线性二次型方法,干扰观测器由比例积分观测器和补偿单元构成,并证明了复合控制器的稳定性。仿真结果表明,将该控制器应用于某型高空无人机非线性模型,可有效减小输油过程带来的影响,实现了输油阶段的位置保持控制。     

System deployment optimization in architecture design

Optimization of architecture design has recently drawn research interest. System deployment optimization （SDO） refers to the process of optimizing systems that are being deployed to activi- ties. This paper first formulates a mathematical model to theorize and operationalize the SDO problem and then identifies optimal so- lutions to solve the SDO problem. In the solutions, the success rate of the combat task is maximized, whereas the execution time of the task and the cost of changes in the system structure are mini- mized. The presented optimized algorithm generates an optimal solution without the need to check the entire search space. A novel method is finally proposed based on the combination of heuristic method and genetic algorithm （HGA）, as well as the combination of heuristic method and particle swarm optimization （HPSO）. Experi- ment results show that the HPSO method generates solutions faster than particle swarm optimization （PSO） and genetic algo- rithm （GA） in terms of execution time and performs more efficiently than the heuristic method in terms of determining the best solution.     

Image separation using wavelet-complex shearlet dictionary

This paper presents a new method for image separation through employing a combined dictionary consisting of wavelets and complex shearlets. Because the combined dictionary sparsely represents points and curvilinear singularities respectively, the image can be decomposed into pointlike and curvelike parts as accurate as possible. The proposed method based on the geo- metric separation theory introduced by Donoho in 2005 shows that accurate geometric separation of the morphologically distinct fea- tures of points and curves can be achieved by l1 minimization. The experimental results show that the proposed method can not only be effective but also greatly reduce the computing time.