Local–class–shared–topic latent Dirichlet allocation based scene classification

In this paper, we propose a hierarchical probabilistic model for scene classification. This model infers the local–class–shared and local–class-specific latent topics respectively. Our approach consists of first learning the latent topics from the BoW representation and subsequently, training SVM on the distribution of the latent topics. This approach is compared to that of using traditional graphical models to learn the latent topics and training SVM on the topic distribution. The experiments on a variety of datasets show that the topics learned by our model have higher discriminative power.     

Pairwise ranking component analysis

Uncovering the latent structure of the data is an active research topic in data mining. However, in the distance metric learning framework, previous studies have mainly focused on the classification performance. In this work, we consider the distance metric learning problem in the ranking setting, where predicting the order between the data vectors is more important than predicting the class labels. We focus on two problems: improving the ranking prediction accuracy and identifying the latent structure of the data. The core of our model consists of ranking the data using a Mahalanobis distance function. The additional use of non-negativity constraints and an entropy-based cost function allows us to simultaneously minimize the ranking error while identifying useful meta-features. To demonstrate its usefulness for information retrieval applications, we compare the performance of our method with four other methods on four UCI data sets, three text data sets, and four image data sets. Our approach shows good ranking accuracies, especially when few training data are available. We also use our model to extract and interpret the latent structure of the data sets. In addition, our approach is simple to implement and computationally efficient and can be used for data embedding and visualization.     

Coupled Action Recognition and Pose Estimation from Multiple Views

Action recognition and pose estimation are two closely related topics in understanding human body movements; information from one task can be leveraged to assist the other, yet the two are often treated separately. We present here a framework for coupled action recognition and pose estimation by formulating pose estimation as an optimization over a set of action-specific manifolds. The framework allows for integration of a 2D appearance-based action recognition system as a prior for 3D pose estimation and for refinement of the action labels using relational pose features based on the extracted 3D poses. Our experiments show that our pose estimation system is able to estimate body poses with high degrees of freedom using very few particles and can achieve state-of-the-art results on the HumanEva-II benchmark. We also thoroughly investigate the impact of pose estimation and action recognition accuracy on each other on the challenging TUM kitchen dataset. We demonstrate not only the feasibility of using extracted 3D poses for action recognition, but also improved performance in comparison to action recognition using low-level appearance features.     

改进在线词对主题模型的微博热点话题演化

话题演化分析是舆情监控的研究热点之一,面向微博热点话题进行演化分析,对于网络用户以及网络监管部门都有很重要的现实意义。针对在线词对主题模型（On-line Biterm Topic Model,OBTM）新旧主题混合、冗余词概率相对较高的问题,对OBTM进行改进,提出基于话题标签和先验参数的OBTM模型（Topic Labels and Prior Parameters OBTM,LPOBTM）。根据微博热点话题的话题标签,将微博文本集区分为含话题标签和不含话题标签的两类数据集,并设置不同的文档-主题先验参数;在前一时间片文档-主题概率分布的基础上,借鉴Sigmod函数对所有主题进行强度排名,从而优化当前时间片上主题-词分布的先验参数计算方法。实验结果表明,LPOBTM能够更准确地描述话题的内容演化情况,并且有更低的模型困惑度。     

基于LSA-HMM的新闻主题分割

主题分割技术是快速并有效地对新闻故事节目进行检索和管理的基础。传统的基于隐马尔可夫模型(HiddenMarkov Model,HMM)的主题分割技术仅使用主题和主题之间的转移寻找主题边界进行新闻分割,并未考虑各主题中词与词之间存在的潜在语义关系。本文提出一种基于隐马尔科夫模型的改进算法。该算法使用潜在语义分析(Latent Se-mantic Analysis,LSA)对词频向量进行特征提取和降维,考虑了词与词之间的上下文关系,通过聚类得到文档类别信息,以LSA特征和主题类别作为HMM的观测和隐状态,这样同时考虑了主题之间的关系,最终实现对文本主题分割。数据实验表明,该算法具有较好的分割性能。     

Automated classification of software change messages by semi-supervised Latent Dirichlet Allocation

ContextTopic models such as probabilistic Latent Semantic Analysis (pLSA) and Latent Dirichlet Allocation (LDA) have demonstrated success in mining software repository tasks. Understanding software change messages described by the unstructured nature-language text is one of the fundamental challenges in mining these messages in repositories.ObjectiveWe seek to present a novel automatic change message classification method characterized by semi-supervised topic semantic analysis.MethodIn this work, we present a semi-supervised LDA based approach to automatically classify change messages. We use domain knowledge of software changes to make labeled samples which are added to build the semi-supervised LDA model. Next, we verify the cross-project analysis application of our method on three open-source projects. Our method has two advantages over existing software change classification methods: First of all, it mitigates the issue of how to set the appropriate number of latent topics. We do not have to choose the number of latent topics in our method, because it corresponds to the number of class labels. Second, this approach utilizes the information provided by the label samples in the training set.ResultsOur method automatically classified about 85% of the change messages in our experiment and our validation survey showed that 70.56% of the time our automatic classification results were in agreement with developer opinions.ConclusionOur approach automatically classifies most of the change messages which record the cause of the software change and the method is applicable to cross-project analysis of software change messages.     

Bidirectional estimation between context and motion in motion sequence in which context changes

ABSTRACT

We describe a motion recognition method for reducing recognition errors. The method has a two-layer structure: a lower layer for motion recognition that is affected by the distribution of topics used as context information in the upper layer and an upper layer for the topic distribution that is affected by motion recognition in the bottom layer. We introduce an algorithm for the bottom layer to integrate the motion likelihood calculated using a hidden Markov model and motion appearance probabilities obtained by a topic model. We also use a set of particles to estimate and update contexts on the basis of the result of motion recognition in the bottom layer. The set of particles presents a probabilistic distribution of motion topics, and motion recognition and particle update procedures are performed on each particle. We experimented with two types of sequential motion: a sequence of 33 daily motions and complex motion sequences assuming actual observation. The results showed that the proposed method reduced recognition errors and tracked latent topics better than conventional methods did.     

A comprehensive system for image scene classification

In recent years, image scene classification based on low/high-level features has been considered as one of the most important and challenging problems faced in image processing research. The high-level features based on semantic concepts present a more accurate and closer model to the human perception of the image scene content. This paper presents a new multi-stage approach for image scene classification based on high-level semantic features extracted from image content. In the first stage, the object boundaries and their labels that represent the content are extracted. For this purpose, a combined method of a fully convolutional deep network and a combined network of a two-class SVM-fuzzy and SVR are used. Topic modeling is used to represent the latent relationships between the objects. Hence in the second stage, a new combination of methods consisting of the bag of visual words, and supervised document neural autoregressive distribution estimator is used to extract the latent topics (topic modeling) in the image. Finally, classification based on Bayesian method is performed according to the extracted features of the deep network, objects labels and the latent topics in the image. The proposed method has been evaluated on three datasets: Scene15, UIUC Sports, and MIT-67 Indoor. The experimental results show that the proposed approach achieves average performance improvement of 12%, 11% and 14% in the accuracy of object detection, and 0.5%, 0.6% and 1.8% in the mean average precision criteria of the image scene classification, compared to the previous state-of-the-art methods on these three datasets.

     

Multi-modal multi-layered topic classification model for social event analysis

In this paper, we pay attention to reveal the event topics and track the evolutionary trend of social event and a novel probabilistic topic model is proposed. The Multi-modal Multi-layered Topic Classification Model (tm_MMC) for Social Event Analysis has the capacity for revealing visual and non-visual topics, by jointly modeling the textual and visual information while simultaneously learning and predicting the multi-layered category labels. In order to track the evolutionary trends of the topics online, tm_MMC uses topic intensity and heritability to incrementally build an up-to-date model. To evaluate the effectiveness of our model, we experiment using a collected data, and compare the results with those of other traditional models. The results demonstrate the effectiveness and advantages of our model against several state-of-the-art methods.     

PSLDA: a novel supervised pseudo document-based topic model for short texts

Various kinds of online social media applications such as Twitter and Weibo, have brought a huge volume of short texts. However, mining semantic topics from short texts efficiently is still a challenging problem because of the sparseness of word-occurrence and the diversity of topics. To address the above problems, we propose a novel supervised pseudo-document-based maximum entropy discrimination latent Dirichlet allocation model (PSLDA for short). Specifically, we first assume that short texts are generated from the normal size latent pseudo documents, and the topic distributions are sampled from the pseudo documents. In this way, the model will reduce the sparseness of word-occurrence and the diversity of topics because it implicitly aggregates short texts to longer and higher-level pseudo documents. To make full use of labeled information in training data, we introduce labels into the model, and further propose a supervised topic model to learn the reasonable distribution of topics. Extensive experiments demonstrate that our proposed method achieves better performance compared with some state-of-the-art methods.     

Synthetic Humans for Action Recognition from Unseen Viewpoints

Although synthetic training data has been shown to be beneficial for tasks such as human pose estimation, its use for RGB human action recognition is relatively unexplored. Our goal in this work is to answer the question whether synthetic humans can improve the performance of human action recognition, with a particular focus on generalization to unseen viewpoints. We make use of the recent advances in monocular 3D human body reconstruction from real action sequences to automatically render synthetic training videos for the action labels. We make the following contributions: (1) we investigate the extent of variations and augmentations that are beneficial to improving performance at new viewpoints. We consider changes in body shape and clothing for individuals, as well as more action relevant augmentations such as non-uniform frame sampling, and interpolating between the motion of individuals performing the same action; (2) We introduce a new data generation methodology, SURREACT, that allows training of spatio-temporal CNNs for action classification; (3) We substantially improve the state-of-the-art action recognition performance on the NTU RGB+D and UESTC standard human action multi-view benchmarks; Finally, (4) we extend the augmentation approach to in-the-wild videos from a subset of the Kinetics dataset to investigate the case when only one-shot training data is available, and demonstrate improvements in this case as well.

     

Gaussian Processes for Object Categorization

Discriminative methods for visual object category recognition are typically non-probabilistic, predicting class labels but not directly providing an estimate of uncertainty. Gaussian Processes (GPs) provide a framework for deriving regression techniques with explicit uncertainty models; we show here how Gaussian Processes with covariance functions defined based on a Pyramid Match Kernel (PMK) can be used for probabilistic object category recognition. Our probabilistic formulation provides a principled way to learn hyperparameters, which we utilize to learn an optimal combination of multiple covariance functions. It also offers confidence estimates at test points, and naturally allows for an active learning paradigm in which points are optimally selected for interactive labeling. We show that with an appropriate combination of kernels a significant boost in classification performance is possible. Further, our experiments indicate the utility of active learning with probabilistic predictive models, especially when the amount of training data labels that may be sought for a category is ultimately very small.     

Unsupervised Learning of Human Action Categories Using Spatial-Temporal Words

We present a novel unsupervised learning method for human action categories. A video sequence is represented as a collection of spatial-temporal words by extracting space-time interest points. The algorithm automatically learns the probability distributions of the spatial-temporal words and the intermediate topics corresponding to human action categories. This is achieved by using latent topic models such as the probabilistic Latent Semantic Analysis (pLSA) model and Latent Dirichlet Allocation (LDA). Our approach can handle noisy feature points arisen from dynamic background and moving cameras due to the application of the probabilistic models. Given a novel video sequence, the algorithm can categorize and localize the human action(s) contained in the video. We test our algorithm on three challenging datasets: the KTH human motion dataset, the Weizmann human action dataset, and a recent dataset of figure skating actions. Our results reflect the promise of such a simple approach. In addition, our algorithm can recognize and localize multiple actions in long and complex video sequences containing multiple motions.     

Cross-domain human action recognition

Conventional human action recognition algorithms cannot work well when the amount of training videos is insufficient. We solve this problem by proposing a transfer topic model (TTM), which utilizes information extracted from videos in the auxiliary domain to assist recognition tasks in the target domain. The TTM is well characterized by two aspects: 1) it uses the bag-of-words model trained from the auxiliary domain to represent videos in the target domain; and 2) it assumes each human action is a mixture of a set of topics and uses the topics learned from the auxiliary domain to regularize the topic estimation in the target domain, wherein the regularization is the summation of Kullback-Leibler divergences between topic pairs of the two domains. The utilization of the auxiliary domain knowledge improves the generalization ability of the learned topic model. Experiments on Weizmann and KTH human action databases suggest the effectiveness of the proposed TTM for cross-domain human action recognition.     

基于概率主题模型的物联网服务发现

针对物联网中服务数量的大规模性、服务描述的异构性以及设备服务的资源高度受限性和移动性等特点,提出了一种基于概率主题模型的物联网服务发现方法.该方法的主要特点是：1） 利用英文Wikipedia 构建高质量的主题模型,并对类似短文本的服务文本描述进行语义扩充,使主题模型能够更有效地估计服务文本描述的隐含主题;2） 提出利用非参数主题模型学习服务文本的隐含主题,降低模型训练时间;3） 利用服务隐含主题对服务进行自动分类和文本相似度计算,快速减少服务匹配数量,加速服务文本相似度计算;4） 提出能够同时支持WSDL-based 和RESTful 两种物联网服务的signature 匹配算法.实验结果表明：与现有的物联网服务发现方法相比,该方法的准确率（precision）和归一化折损累积增益（NDCG）都有较大幅度的提高.     

A two-level model for automatic image annotation

Image automatic annotation is a significant and challenging problem in pattern recognition and computer vision. Current image annotation models almost used all the training images to estimate joint generation probabilities between images and keywords, which would inevitably bring a lot of irrelevant images. To solve the above problem, we propose a hierarchical image annotation model which combines advantages of discriminative model and generative model. In first annotation layer, discriminative model is used to assign topic annotations to unlabeled images, and then relevant image set corresponding to each unlabeled image is obtained. In second annotation layer, we propose a keywords-oriented method to establish links between images and keywords, and then our iterative algorithm is used to expand relevant image sets. Candidate labels will be given higher weights by using our method based on visual keywords. Finally, generative model is used to assign detailed annotations to unlabeled images on expanded relevant image sets. Experiments conducted on Corel 5K datasets verify the effectiveness of our hierarchical image annotation model.     

融合语义主题的图像自动标注

由于语义鸿沟的存在,图像自动标注已成为一个重要课题.在概率潜语义分析的基础上,提出了一种融合语义主题的方法以进行图像的标注和检索.首先,为了更准确地建模训练数据,将每幅图像的视觉特征表示为一个视觉"词袋";然后设计一个概率模型分别从视觉模态和文本模态中捕获潜在语义主题,并提出一种自适应的不对称学习方法融合两种语义主题.对于每个图像文档,它在各个模态上的主题分布通过加权进行融合,而权值由该文档的视觉词分布的熵值来确定.于是,融合之后的概率模型适当地关联了视觉模态和文本模态的信息,因此能够很好地预测未知图像的语义标注.在一个通用的Corel图像数据集上,将提出的方法与几种前沿的图像标注方法进行了比较.实验结果表明,该方法具有更好的标注和检索性能.     

Automated topic naming

Software repositories provide a deluge of software artifacts to analyze. Researchers have attempted to summarize, categorize, and relate these artifacts by using semi-unsupervised machine-learning algorithms, such as Latent Dirichlet Allocation (LDA). LDA is used for concept and topic analysis to suggest candidate word-lists or topics that describe and relate software artifacts. However, these word-lists and topics are difficult to interpret in the absence of meaningful summary labels. Current attempts to interpret topics assume manual labelling and do not use domain-specific knowledge to improve, contextualize, or describe results for the developers. We propose a solution: automated labelled topic extraction. Topics are extracted using LDA from commit-log comments recovered from source control systems. These topics are given labels from a generalizable cross-project taxonomy, consisting of non-functional requirements. Our approach was evaluated with experiments and case studies on three large-scale Relational Database Management System (RDBMS) projects: MySQL, PostgreSQL and MaxDB. The case studies show that labelled topic extraction can produce appropriate, context-sensitive labels that are relevant to these projects, and provide fresh insight into their evolving software development activities.