基于深度学习和三支决策的DDoS攻击检测算法

针对软件定义网络(Software Defined Ntwork,SDN)中的分布式拒绝服务(Distribute Denial of Service,DDoS)攻击检测的方法少、现存方法入侵检测率低的问题,提出了一种基于深度学习和三支决策的入侵检测算法.首先使用深度信念网络对SDN的流表项进行特征提取,然后利用基于三支决策理论的入侵检测模型进行DDoS攻击的入侵检测,对于正域和负域的数据直接进行分类,对于边界域中的数据使用K近邻算法重新进行分类.仿真实验结果表明,与其他入侵检测模型相比,所提算法的入侵检测效率更高.     

引入源端信息的机器译文自动评价方法研究

机器译文自动评价是机器翻译中的一个重要任务。针对目前译文自动评价中完全忽略源语言句子信息,仅利用人工参考译文度量翻译质量的不足,该文提出了引入源语言句子信息的机器译文自动评价方法: 从机器译文与其源语言句子组成的二元组中提取描述翻译质量的质量向量,并将其与基于语境词向量的译文自动评价方法利用深度神经网络进行融合。在WMT-19译文自动评价任务数据集上的实验结果表明,该文所提出的方法能有效增强机器译文自动评价与人工评价的相关性。深入的实验分析进一步揭示了源语言句子信息在译文自动评价中发挥着重要作用。     

一种基于上下文信息的个性化推荐模型

个性化推荐为解决互联网信息过载问题提供了新的思路。为有效地构建用户模型和改进个性化推荐的效果,提出了一种挖掘非结构化文本中上下文信息的新模型,将得到的上下文信息嵌入用户模型信息中,丰富了用户模型。实验结果表明,该模型应用于客户对旅馆评论的上下文数据中,能够大大改善推荐的性能。     

Harmonizing architectural decisions with component view models using reusable architectural knowledge transformations and constraints

Architectural design decisions (ADDs) have been used in recent years for capturing design rationale and documenting architectural knowledge (AK). However, various architectural design views still provide the most common means for describing and communicating architectural design. The evolution of software systems requires that both ADDs and architectural design views are documented and maintained, which is a tedious and time-consuming task in the long run. Also, in lack of a systematic and automated support for bridging between ADDs and architectural design views, decisions and designs tend to become inconsistent over time. In our proposal, we introduce a reusable AK transformation language for supporting the automated transformation of reusable AK knowledge to component-and-connector models, the architectural design view used most commonly today. In addition, reusable consistency checking rules verify the consistency between decisions and designs. We evaluate our approach in an industrial case study and show that it offers high reusability, provides automation, and can, in principle, deal with large numbers of recurring decisions.     

基于网络图的产品生产、运输、定价综合决策模型

针对大型垄断厂商在商品生产、运输、定价决策中独立考虑生产定价和运输两个环节的弊端, 提出一种将产量决策、运输决策和定价决策3 个环节作为一个系统进行考虑的综合决策模型. 借助利润网络图, 构建同时考虑生产及运输成本以及市场需求的商品生产、运输和定价综合决策模型, 并探讨模型的解法及其理论基础. 通过给出的应用案例表明, 相比传统决策模型, 利用综合决策模型可以提高15.50% 的利润.

     

蚁群遗传算法的多目标优化

为了求解带有约束条件的多目标函数优化问题,提出基于连续空间优化的多目标蚁群遗传算法。针对多目标优化问题的特点,定义连续空间中利用信息量指导遗传搜索策略和信息更新方法,将信息量指导遗传搜索、优秀决策引入、决策集更新、改变算法终止条件等方式相结合,有效地加速了搜索的收敛速度,控制了Pareto最优决策集的数量,扩大了决策的分布范围,维持了决策的多样性。数值实验说明该算法能够快速找到一组分布广泛的Pareto最优决策。     

关于扩充转移网络的研究

随着因特网的普及,计算机整体性能已经大幅提高,信息处理量也出现爆炸式增长,在这种情况下,要求计算机对我们的母语（自然语言）理解日益迫切。自然语言理解的主流技术一直是句法--语义分析。本文研究了扩充转移网络的描述语言的语法,给出了扩充转移网络的分析算法,并通过实例解析了这种算法的特点。     

基于改进三角模糊VIKOR法的用户储能系统配置方案综合评估

针对不确定环境下用户储能系统配置方案评估决策,提出一种基于改进三角模糊多准则妥协解排序(VIKOR)法的评估模型。首先,运用扎根理论构建评估指标集,并采用灰色关联-粗糙集方法进行指标筛选,形成评估指标体系。然后,采用主、客观相结合的方法计算指标组合权重,以提高其准确性,其中主观权重由模糊层次分析(FAHP)法求得,客观赋权法由改进标准间重要性相关性(CRITIC)法求得。最后,考虑指标之间的冲突性与模糊性,运用改进三角模糊VIKOR法对配置方案评估与排序比选。通过仿真分析,验证了本文所建模型的有效性和可行性。     

Residual Feature Attentional Fusion Network for Lightweight Chest CT Image Super-Resolution

The diagnosis of COVID-19 requires chest computed tomography (CT). High-resolution CT images can provide more diagnostic information to help doctors better diagnose the disease, so it is of clinical importance to study super-resolution (SR) algorithms applied to CT images to improve the resolution of CT images. However, most of the existing SR algorithms are studied based on natural images, which are not suitable for medical images; and most of these algorithms improve the reconstruction quality by increasing the network depth, which is not suitable for machines with limited resources. To alleviate these issues, we propose a residual feature attentional fusion network for lightweight chest CT image super-resolution (RFAFN). Specifically, we design a contextual feature extraction block (CFEB) that can extract CT image features more efficiently and accurately than ordinary residual blocks. In addition, we propose a feature-weighted cascading strategy (FWCS) based on attentional feature fusion blocks (AFFB) to utilize the high-frequency detail information extracted by CFEB as much as possible via selectively fusing adjacent level feature information. Finally, we suggest a global hierarchical feature fusion strategy (GHFFS), which can utilize the hierarchical features more effectively than dense concatenation by progressively aggregating the feature information at various levels. Numerous experiments show that our method performs better than most of the state-of-the-art (SOTA) methods on the COVID-19 chest CT dataset. In detail, the peak signal-to-noise ratio (PSNR) is 0.11 dB and 0.47 dB higher on CTtest1 and CTtest2 at SR compared to the suboptimal method, but the number of parameters and multi-adds are reduced by 22K and 0.43G, respectively. Our method can better recover chest CT image quality with fewer computational resources and effectively assist in COVID-19.