基于深度学习的GPS轨迹去匿名研究

移动互联网和LBS技术的高速发展使得位置服务提供商可以轻松收集到大量用户位置轨迹数据,近期研究表明,深度学习方法能够从轨迹数据集中提取出用户身份标识等隐私信息.然而现有工作主要针对社交网络采集的签到点轨迹,针对GPS轨迹的去匿名研究则较为缺乏.因此,对基于深度学习的GPS轨迹去匿名技术开展研究.首先提出一种GPS轨迹数...     

无人机辅助的双层深度强化学习任务卸载算法

为了解决无人机轨迹优化、用户功率分配和任务卸载策略问题,提出了一种双层深度强化学习任务卸载算法。上层采用多智能体深度强化学习来优化无人机的轨迹,并动态分配用户的传输功率以提高网络传输速率;下层采用多个并行的深度神经网络来求解最优卸载决策以最小化网络的时延和能耗。仿真结果表明,该算法使得无人机能够跟踪用户的移动,显著降低系统的时延和能耗,能够给用户提供更优质的任务卸载服务。     

基于深度学习的在线健康社区抑郁症用户画像研究

人工智能与深度学习技术为精准识别在线健康社区抑郁症患者奠定了基础.首先构建了基于TCNN-GRU深度学习的抑郁情感分类模型,进行在线健康社区实验数据集进行抑郁情感分类标注后,通过TCNN-GRU模型判别用户的抑郁症倾向;在此基础上,进一步提出抑郁指数的概念,通过对抑郁指数和患者抑郁程度两者关系的深度挖掘,由此建立基于深度学习的在线健康社区抑郁症用户画像模型.实验结果表明,与传统的卷积神经网络模型、循环神经网络模型以及混合模型相比,TCNN-GRU模型在抑郁情感分类上能获得了更优的结果,基于深度学习的在线健康社区抑郁症用户画像模型也能够从文本分析的角度准确识别用户的抑郁情感和抑郁状态.     

Dynamic classification of ballistic missiles using neural networks and hidden Markov models

This paper addresses dynamic classification of different ranges of ballistic missiles (BM) for air defense application based on kinematic attributes acquired by radars for taking appropriate measures to intercept them. The problem of dynamic classification is formulated using real-time neural network (RTNN) and hidden Markov model (HMM). The idea behind these algorithms is to calculate the output in one pass rather than training and computing over large number of iterations. Besides, to meet the conflicting requirements of classifying small as well as long-range trajectories, we are also proposing a formulation for partitioning the trajectory by using moving window concept. This concept allows us to use parameters in localized frame which helps in handling wide-range of trajectories to fit into the same network. These algorithms are evaluated using the simulated data generated from 6 degree-of-freedom (6DOF) mathematical model, which models missile trajectories. Experimental results show that both the networks are classifying above 95% with real-time neural network outperforming HMM in terms of time of computation on same data. The small classification time enables the use of real-time classification neural network in complex scenario of multi-radar, multi-target engagement by interceptor missiles. To the best of our knowledge this is the first time an attempt is made to classify ballistic missiles using RTNN and HMM.     

轨迹表示学习技术研究进展

基于地理位置信息的应用和服务的迅速发展对轨迹数据挖掘提出新的需求和挑战.原始轨迹数据通常是由坐标-时间戳元组构成的有序序列组成,而现有的大多数数据分析算法均要求输入数据位于向量空间中.因此,为了将轨迹数据从变长的坐标-时间戳序列转化定长的向量表示且保持原有的特征,对轨迹数据进行有效的表示是十分重要且必要的一步.传统的轨迹表示方法多是基于人工设计特征,通常仅将轨迹表示作为数据预处理的一部分.随着深度学习的兴起,这种从大规模数据中学习的能力使得基于深度学习的轨迹表示方法相较于传统方法取得了巨大的效果提升,并赋予了轨迹表示更多的可能性.本文对轨迹表示领域中的研究进展进行了全面的总结,将轨迹表示按照研究对象的不同尺度归纳为对轨迹单元的表示和对整条轨迹的表示两大类别,并在每种类别下对不同原理的方法进行了对比分析.其中重点分析了基于轨迹点表示的关键方法,也对近年来广泛使用的基于神经网络的轨迹表示的研究成果做了系统的归类.此外本文介绍了基于轨迹表示的关键应用,最后对轨迹表示领域的未来研究方向进行了展望.     

Deep learning with particle filter for person re-identification

Person re-identification, having attracted much attention in the multimedia community, is still challenged by the accuracy and the robustness, as the images for the verification contain such variations as light, pose, noise and ambiguity etc. Such practical challenges require relatively robust and accurate feature learning technologies. We introduced a novel deep neural network with PF-BP(Particle Filter-Back Propagation) to achieve relatively global and robust performances of person re-identification. The local optima in the deep networks themselves are still the main difficulty in the learning, in despite of several advanced approaches. A novel neural network learning, or PF-BP, was first proposed to solve the local optima problem in the non-convex objective function of the deep networks. When considering final deep network to learn using BP, the overall neural network with the particle filter will behave as the PF-BP neural network. Also, a max-min value searching was proposed by considering two assumptions about shapes of the non-convex objective function to learn on. Finally, a salience learning based on the deep neural network with PF-BP was proposed to achieve an advanced person re-identification. We test our neural network learning with particle filter aimed to the non-convex optimization problem, and then evaluate the performances of the proposed system in a person re-identification scenario. Experimental results demonstrate that the corresponding performances of the proposed deep network have promising discriminative capability in comparison with other ones.

     

基于高速公路场景的车辆目标跟踪

车辆目标检测与跟踪是高速公路视频监控系统实时监控获取交通参数的关键步骤.本文提出了一种面向高速公路场景的目标轨迹时序信息结合核相关滤波KCF算法的车辆目标跟踪方法,实现了车辆目标的高精度持续跟踪.该方法首先采用基于深度学习的单目标检测SSD算法,通过建立车辆数据集,实现了适用于高速公路场景的车辆目标的分类与检测.然后,基于目标轨迹时序信息实现目标车辆与轨迹的匹配,并且采用KCF跟踪算法对丢失目标进行预测重定位,从而实现车辆目标轨迹的持续跟踪.实验表明,该跟踪方法精度高,且适应多种不同场景,具有较高的应用价值.     

Design and development of an indoor navigation system using denoising autoencoder based convolutional neural network for visually impaired people

A challenging area of research is the development of a navigation system for visually impaired people in an indoor environment such as a railway station, commercial complex, educational institution, and airport. Identifying the current location of the users can be a difficult task for those with visual impairments. The entire selection of the navigation path depends upon the current location of the user. This work presents a detailed analysis of the recent user positioning techniques and methodologies on the indoor navigation system based on the parameters, such as techniques, cost, the feasibility of implementation, and limitations. This paper presents a denoising auto encoder based on the convolutional neural network (DAECNN) to identify the present location of the users. The proposed approach uses the de-noising autoencoder to reconstruct the noisy image and the convolution neural network (CNN) to classify the users' current position. The proposed method is compared with the existing deep learning approaches such as deep autoencoder, sparse autoencoder, CNN, multilayer perceptron, radial basis function neural network, and the performances are analyzed. The experimental findings indicate that the DAECNN methodology works better than the existing classification approaches.

     

A survey on deep learning-based fine-grained object classification and semantic segmentation

The deep learning technology has shown impressive performance in various vision tasks such as image classification, object detection and semantic segmentation. In particular, recent advances of deep learning techniques bring encouraging performance to fine-grained image classification which aims to distinguish subordinate-level categories, such as bird species or dog breeds. This task is extremely challenging due to high intra-class and low inter-class variance. In this paper, we review four types of deep learning based fine-grained image classification approaches, including the general convolutional neural networks (CNNs), part detection based, ensemble of networks based and visual attention based fine-grained image classification approaches. Besides, the deep learning based semantic segmentation approaches are also covered in this paper. The region proposal based and fully convolutional networks based approaches for semantic segmentation are introduced respectively.     

基于改进SVM的互联网用户分类

由于传统模型大量约束样本,导致其学习能力下降,因此设计一个基于改进支持向量机(Support Vector Machine,SVM)的互联网用户分类模型.该模型通过构造样本数据,模拟互联网用户的浏览轨迹;根据用户偏好,制定全新的用户分类策略;基于改进支持向量机,实现对互联网用户的分类.性能测试:3次实验下,此次设计的模型分类准确率平均值为98.56%,超出了预设的期望值,具备分类能力.对比测试:与两组传统用户分类模型相比,此次设计的模型,面对不断增加的样本数据,同样能保持高水平的学习能力.     

Incremental deep forest for multi-label data streams learning

Multi-label learning has attracted many attentions. However, the continuous data generated in the fields of sensors, network access, etc., that is data streams, the scenario brings challenges such as real-time, limited memory, once pass. Several learning algorithms have been proposed for offline multi-label classification, but few researches develop it for dynamic multi-label incremental learning models based on cascading schemes. Deep forest can perform representation learning layer by layer, and does not rely on backpropagation, using this cascading scheme, this paper proposes a multi-label data stream deep forest (VDSDF) learning algorithm based on cascaded Very Fast Decision Tree (VFDT) forest, which can receive examples successively, perform incremental learning, and adapt to concept drift. Experimental results show that the proposed VDSDF algorithm, as an incremental classification algorithm, is more competitive than batch classification algorithms on multiple indicators. Moreover, in dynamic flow scenarios, the adaptability of VDSDF to concept drift is better than that of the contrast algorithm.

     

基于轨迹时空词向量的用户年龄特征识别

用户移动上网访问基站的轨迹数据从时间和空间上反映了用户的生活习惯和行为模式。时间和空间信息同时产生不应分别考虑。因此,该文在传统的TF-IDF方法基础上提出了与时间相关的TFT-IDFT方法,用以提取轨迹点语义信息,进而采用word2vec方法将轨迹数据转化为文档分析。提取包含位置信息和语义信息的轨迹时空词向量,在此基础上建立多分类模型对用户所属年龄段进行识别。实验结果表明,改进的TFT-IDFT方法在提取轨迹语义时更具合理性,且基于此方法构建的轨迹时空词向量应用于分类模型,对用户所属年龄阶段的识别效果更好。     

深度学习方法在兴趣点推荐中的应用研究综述

在基于位置的社交网络（LBSN）中,用户可以在兴趣点（POI）进行签到以记录行程,也可以与其他用户分享自身的感受并形成社交好友关系。POI推荐是LBSN提供的一项重要服务,其可以帮助用户快速发现感兴趣的POI,也有利于POI提供商更全面地了解用户偏好,并有针对性地提高服务质量。POI推荐主要基于对用户历史签到数据以及用户生成内容、社交关系等信息的分析来实现。系统归纳POI推荐中所面临的时空序列特征提取、内容社交特征提取、多特征整合、数据稀疏性问题处理这4个方面的挑战,分析在POI推荐中使用深度学习方法解决上述问题时存在的优势以及不足。在此基础上,展望未来通过深度学习提高POI推荐效果的研究方向,即通过增量学习加速推荐模型更新、使用迁移学习缓解冷启动问题以及利用强化学习建模用户动态偏好,从而为实现效率更高、用户体验质量更好的推荐系统提供新的思路。     

Deep infrared pedestrian classification based on automatic image matting

Infrared pedestrian classification plays an important role in advanced driver assistance systems. However, it encounters great difficulties when the pedestrian images are superimposed on a cluttered background. Many researchers design very deep neural networks to classify pedestrian from cluttered background. However, a very deep neural network associated with a high computational cost. The suppression of cluttered background can boost the performance of deep neural networks without increasing their depth, while it has received little attention in the past. This study presents an automatic image matting approach for infrared pedestrians that suppresses the cluttered background and provides consistent input to deep learning. The domain expertise in pedestrian classification is applied to automatically and softly extract foreground objects from images with cluttered backgrounds. This study generates trimaps, which must be generated manually in conventional approaches, according to the estimated positions of pedestrian’s head and upper body without the need for any user interaction. We implement image matting by adopting the global matting approach and taking the generated trimap as an input. The representation of pedestrian is discovered by a deep learning approach from the resulting alpha mattes in which cluttered background is suppressed, and foreground is enhanced. The experimental results show that the proposed approach improves the infrared pedestrian classification performance of the state-of-the-art deep learning approaches at a negligible computational cost.     

Supervised machine learning in multimodal learning analytics for estimating success in project‐based learning

Multimodal learning analytics provides researchers new tools and techniques to capture different types of data from complex learning activities in dynamic learning environments. This paper investigates the use of diverse sensors, including computer vision, user‐generated content, and data from the learning objects (physical computing components), to record high‐fidelity synchronised multimodal recordings of small groups of learners interacting. We processed and extracted different aspects of the students' interactions to answer the following question: Which features of student group work are good predictors of team success in open‐ended tasks with physical computing? To answer this question, we have explored different supervised machine learning approaches (traditional and deep learning techniques) to analyse the data coming from multiple sources. The results illustrate that state‐of‐the‐art computational techniques can be used to generate insights into the "black box" of learning in students' project‐based activities. The features identified from the analysis show that distance between learners' hands and faces is a strong predictor of students' artefact quality, which can indicate the value of student collaboration. Our research shows that new and promising approaches such as neural networks, and more traditional regression approaches can both be used to classify multimodal learning analytics data, and both have advantages and disadvantages depending on the research questions and contexts being investigated. The work presented here is a significant contribution towards developing techniques to automatically identify the key aspects of students success in project‐based learning environments, and to ultimately help teachers provide appropriate and timely support to students in these fundamental aspects.     

医学图像分析深度学习方法研究与挑战

深度学习（Deep learning,DL）,特别是深度卷积神经网络（Convolutional neural networks,CNNs）,能够从医学图像大数据中自动学习提取隐含的疾病诊断特征,近几年已迅速成为医学图像分析研究热点.本文首先简述医学图像分析特点;其次,论述深度学习基本原理,总结深度CNNs在医学图像分析中的分类、分割框架;然后,分别论述深度学习在医学图像分类、检测、分割等各应用领域的国内外研究现状;最后,探讨归纳医学图像分析深度学习方法挑战及其主要应对策略和开放的研究方向.     

Learning to rank biological motion trajectories

Many feature transforms have been proposed for the problem of trajectory matching. These methods, which are often based on shape matching, tend to perform poorly for biological trajectories, such as cell motion, because similar biological behavior often results in dissimilar trajectory shape. Additionally, the criteria used for similarity may differ depending on the user's particular interest or the specific query behavior. We present a rank-based distance metric learning method that combines user input and a new set of biologically-motivated features for biological trajectory matching. We show that, with a small amount of user effort, this method outperforms existing trajectory methods. On an information retrieval task using real world data, our method outperforms recent, related methods by ~ 9%.     

基于自监督学习的社交网络用户轨迹预测模型

针对当前用户轨迹数据建模中存在的签到点稀疏性、长时间依赖性和移动模式复杂等问题,提出基于自监督学习的社交网络用户轨迹预测模型SeNext,对用户轨迹进行建模和训练来预测用户的下一个兴趣点（POI）。首先,使用数据增强的方式来丰富训练数据样本,以解决数据不足及个别用户足迹太少导致的模型泛化能力不足的问题;其次,将循环神经网络（RNN）、卷积神经网络（CNN）和注意力机制分别用于当前轨迹和历史轨迹的建模中,以此从高维稀疏的数据中提取有用的表示,用来匹配用户过去最相似的移动方式。最后,通过结合自监督学习并引入对比损失优化噪声对比估计（InfoNCE）,SeNext在潜在空间学习隐含表示来预测用户的下一个POI。实验结果表明,在纽约数据集上,SeNext比最新的VANext（Variational Attention based Next）模型的预测准确度在Top@1上提高了11.10%左右。