基于深度极限学习机的危险源识别算法HIELM

危险源识别是民用航空管理的重要环节之一,危险源识别结果必须高度准确才能确保飞行的安全。为此,提出了一种基于深度极限学习机的危险源识别算法HIELM(Hazard Identification Algorithm Based on Extreme Lear-ning Machine),设计了一种由多个深层栈式极限学习机(S-ELM)和一个单隐藏层极限学习机(ELM)构成的深层网络结构。算法中,多个深层S-ELM使用平行结构,各自可以拥有不同的隐藏结点个数,按照危险源领域分类接受危险源状态信息完成预学习,并结合识别特征改进网络输入权重的产生方式。在单隐藏层ELM中,深层ELM的预学习结果作为其输入,改进了反向传播算法,提高了网络识别的精确度。同时,分别训练各深层S-ELM,缓解了高维数据训练的内存压力和节点过多产生的过拟合现象。     

基于SDAs的人物关系抽取方法研究

针对人物关系语料缺乏的问题,研究了基于互动百科的自动标注方法;针对传统浅层机器学习模型特征表示能力差的问题,提出了基于深度神经网络模型SDAs的人物关系抽取方法。重点研究了多个特征组合的人物关系抽取效果以及不同深度SDAs网络的人物关系抽取效果。根据实验分析,F系数可达到73.75%。     

Some Challenges of Deep Mining

An increased global supply of minerals is essential to meet the needs and expectations of a rapidly rising world population. This implies extraction from greater depths. Autonomous mining systems, developed through sustained R&D by equipment suppliers, reduce miner exposure to hostile work environments and increase safety. This places increased focus on “ground control” and on rock mechanics to define the depth to which minerals may be extracted economically. Although significant efforts have been made since the end of World War II to apply mechanics to mine design, there have been both technological and organizational obstacles. Rock in situ is a more complex engineering material than is typically encountered in most other engineering disciplines. Mining engineering has relied heavily on empirical procedures in design for thousands of years. These are no longer adequate to address the challenges of the 21st century, as mines venture to increasingly greater depths. The development of the synthetic rock mass (SRM) in 2008 provides researchers with the ability to analyze the deformational behavior of rock masses that are anisotropic and discontinuous—attributes that were described as the defining characteristics of in situ rock by Leopold Müller, the president and founder of the International Society for Rock Mechanics (ISRM), in 1966. Recent developments in the numerical modeling of large-scale mining operations (e.g., caving) using the SRM reveal unanticipated deformational behavior of the rock. The application of massive parallelization and cloud computational techniques offers major opportunities: for example, to assess uncertainties in numerical predictions; to establish the mechanics basis for the empirical rules now used in rock engineering and their validity for the prediction of rock mass behavior beyond current experience; and to use the discrete element method (DEM) in the optimization of deep mine design. For the first time, mining—and rock engineering—will have its own mechanics-based “laboratory.” This promises to be a major tool in future planning for effective mining at depth. The paper concludes with a discussion of an opportunity to demonstrate the application of DEM and SRM procedures as a laboratory, by back-analysis of mining methods used over the 80-year history of the Mount Lyell Copper Mine in Tasmania.     

Parallel inverter control using different conventional control methods and an improved virtual oscillator control method in a standalone microgrid

This paper develops a multi-timescale coordinated operation method for microgrids based on modern deep reinforcement learning. Considering the complementary characteristics of different storage devices, the proposed approach achieves multi-timescale coordination of battery and supercapacitor by introducing a hierarchical two-stage dispatch model. The first stage makes an initial decision irrespective of the uncertainties using the hourly predicted data to minimize the operational cost. For the second stage, it aims to generate corrective actions for the first-stage decisions to compensate for real-time renewable generation fluctuations. The first stage is formulated as a non-convex deterministic optimization problem, while the second stage is modeled as a Markov decision process solved by an entropy-regularized deep reinforcement learning method, i.e., the Soft Actor-Critic. The Soft Actor-Critic method can efficiently address the exploration–exploitation dilemma and suppress variations. This improves the robustness of decisions. Simulation results demonstrate that different types of energy storage devices can be used at two stages to achieve the multi-timescale coordinated operation. This proves the effectiveness of the proposed method.     

Profile and Rating Similarity Analysis for Recommendation Systems Using Deep Learning

Recommendation systems are going to be an integral part of any E-Business in near future. As in any other E-business, recommendation systems also play a key role in the travel business where the user has to be recommended with a restaurant that best suits him. In general, the recommendations to a user are made based on similarity that exists between the intended user and the other users. This similarity can be calculated either based on the similarity between the user profiles or the similarity between the ratings made by the users. First phase of this work concentrates on experimentally analyzing both these models and get a deep insight of these models. With the lessons learned from the insights, second phase of the work concentrates on developing a deep learning model. The model does not depend on the other user's profile or rating made by them. The model is tested with a small restaurant dataset and the model can predict whether a user likes the restaurant or not. The model is trained with different users and their rating. The system learns from it and in order to predict whether a new user likes or not a restaurant that he/she has not visited earlier, all the data the trained model needed is the rating made by the same user for different restaurants. The model is deployed in a cloud environment in order to extend it to be more realistic product in future. Result evaluated with dataset, it achieves 74.6% is accurate prediction of results, where as existing techniques achieves only 64%.     

基于孪生区域建议网络的无人机目标跟踪算法

在无人机跟踪过程中,遮挡、光照变化、背景干扰等影响会导致跟踪目标丢失。基于SiamRPN算法提出一种无人机目标跟踪算法。通过在网络中加入空间条带池和全局上下文模块建立远程上下文关系,以适应不同的跟踪场景。同时利用改进交并比的计算方法提取目标特征,并回归精准的预测框。在UAV123数据集上的实验结果表明,相比SiamRPN、SiamFC、SAMF等算法,该算法的跟踪性能较优且具有较强的鲁棒性,尤其在背景干扰环境下,其精确率和成功率较SiamRPN算法分别提升了6.54%和11.63%。     

Effective short text classification via the fusion of hybrid features for IoT social data

Nowadays short texts can be widely found in various social data in relation to the 5G-enabled Internet of Things (IoT). Short text classification is a challenging task due to its sparsity and the lack of context. Previous studies mainly tackle these problems by enhancing the semantic information or the statistical information individually. However, the improvement achieved by a single type of information is limited, while fusing various information may help to improve the classification accuracy more effectively. To fuse various information for short text classification, this article proposes a feature fusion method that integrates the statistical feature and the comprehensive semantic feature together by using the weighting mechanism and deep learning models. In the proposed method, we apply Bidirectional Encoder Representations from Transformers (BERT) to generate word vectors on the sentence level automatically, and then obtain the statistical feature, the local semantic feature and the overall semantic feature using Term Frequency-Inverse Document Frequency (TF-IDF) weighting approach, Convolutional Neural Network (CNN) and Bidirectional Gate Recurrent Unit (BiGRU). Then, the fusion feature is accordingly obtained for classification. Experiments are conducted on five popular short text classification datasets and a 5G-enabled IoT social dataset and the results show that our proposed method effectively improves the classification performance.     

Moving target defense of routing randomization with deep reinforcement learning against eavesdropping attack

Eavesdropping attacks have become one of the most common attacks on networks because of their easy implementation. Eavesdropping attacks not only lead to transmission data leakage but also develop into other more harmful attacks. Routing randomization is a relevant research direction for moving target defense, which has been proven to be an effective method to resist eavesdropping attacks. To counter eavesdropping attacks, in this study, we analyzed the existing routing randomization methods and found that their security and usability need to be further improved. According to the characteristics of eavesdropping attacks, which are “latent and transferable”, a routing randomization defense method based on deep reinforcement learning is proposed. The proposed method realizes routing randomization on packet-level granularity using programmable switches. To improve the security and quality of service of legitimate services in networks, we use the deep deterministic policy gradient to generate random routing schemes with support from powerful network state awareness. In-band network telemetry provides real-time, accurate, and comprehensive network state awareness for the proposed method. Various experiments show that compared with other typical routing randomization defense methods, the proposed method has obvious advantages in security and usability against eavesdropping attacks.     

基于深度神经网络的社会媒体网络分析

社会媒体网络中不仅包含了用户、文本、图片和视频等多种模态的数据,还包含了反映不同模态数据之间交互的群体特征。为了更好地描述社会媒体网络,从而为上层应用提供更好的服务,提出了一种基于深度神经网络的社会媒体网络模型。该模型采用深度神经网络对单个模态的数据进行学习,从而得到任意一个模态数据的潜在特征表示方法。对于两种不同模态的数据,利用具有高斯分布的先验矩阵与两个模态数据的后验分布建立反映这两个模态数据间群体特征的生成模型。实验结果表明,提出的模型在网络结构的链接分析中具有更好的预测效果,能有效地描述社会媒体网络的整体特征。     

基于耦合关系模型的文本分类研究

对文本的特征提取方法以及深度神经网络的分类器的搭建进行研究。首先,在全局和局部的特征提取方法的基础上,通过对文本特征内耦合关系和文本特征间耦合关系进行分析,确定用于分类的文本特征,建立文本特征的耦合关系模型;其次, 将文本特征 作为深度神经网络输入层进行分类;最后,通过逐层无监督的方式对网络进行训练,在顶层增加区分性结点来实现文本分类功能。