Probabilistic Graph Research Articles

Knowledge Graph Completion: A Review

Knowledge graph completion (KGC) is a hot topic in knowledge graph construction and related applications, which aims to complete the structure of knowledge graph by predicting the missing entities or relationships in knowledge graph and mining unknown facts. Starting from the definition and types of KGC, existing technologies for KGC are analyzed in categories. From the evolving point of view, the KGC technologies could be divided into traditional and representation learning based methods. The former mainly includes rule-based reasoning method, probability graph model, such as Markov logic network, and graph computation based method. The latter further includes translation model based, semantic matching model based, representation learning based and other neural network model based methods. In this article, different KGC technologies are introduced, including their advantages, disadvantages and applicable fields. Finally the main challenges and problems faced by the KGC are discussed, as well as the potential research directions.

Sparse and Low-Rank Representation With Key Connectivity for Hyperspectral Image Classification

Combined techniques of sparse representation (SR) and low-rank representation (LRR) are commonly used for hyperspectral image (HSI) classification. Although they have the ability to capture the interclass representations of data for HSI classification, they ignore the adaptive key connectivity of the learned intraclass data representations in particular with the high-dimensional complex HSI data. It is well-known that the key connectivity of graph-based algorithms is crucial for subspace learning because of the guarantees of its good neighbors. For this purpose, a novel sparse and low-rank representation with key connectivity (SLRC) method is proposed for HSI classification. To be specific, the adaptive probability graph structure is developed to integrate the SR and LRR regularizations to formulate the SLRC model, which flexibly perform discriminative latent subspace construction and preserve the key connectivity of intraclass representations. Then, extensive experiments are executed based on three popular HSI datasets, which demonstrates that the SLRC method outperforms the other popular methods.

Probabilistic Kalman filter for moving object tracking

Kalman filter has been successfully applied to tracking moving objects in real-time situations. However, the filter cannot take into account the existing prior knowledge to improve its predictions. In the moving object tracking, the trajectories of multiple targets in the same environment could be available, which can be viewed as the prior knowledge for the tracking procedure. This paper presents the probabilistic Kalman filter (PKF) that is able to take into account the stored trajectories to improve tracking estimation. The PKF has an extra stage after two steps of the Kalman filter to refine the estimated position of the targets. The refinement is obtained by applying the Viterbi algorithm to a probabilistic graph, that is constructed based on the observed trajectories. The graph is built in the offline situation and could be adapted in the online tracking. The proposed tracker has higher accuracy compared to the standard Kalman filter and could handle widespread problems such as occlusion. Another significant achievement of the proposed tracker is to track an object with anomalous behaviors by drawing an inference based on the constructed probabilistic graph. The PKF was applied to several manually-built videos and several other video-bases containing severe occlusions, which demonstrates a significant performance in comparison with other state-of-the-art trackers.

Heterogeneous information network based clustering for precision traditional Chinese medicine

BackgroundTraditional Chinese medicine (TCM) is a highly important complement to modern medicine and is widely practiced in China and in many other countries. The work of Chinese medicine is subject to the two factors of the inheritance and development of clinical experience of famous Chinese medicine practitioners and the difficulty in improving the service capacity of basic Chinese medicine practitioners. Heterogeneous information networks (HINs) are a kind of graphical model for integrating and modeling real-world information. Through HINs, we can integrate and model the large-scale heterogeneous TCM data into structured graph data and use this as a basis for analysis.MethodsMining categorizations from TCM data is an important task for precision medicine. In this paper, we propose a novel structured learning model to solve the problem of formula regularity, a pivotal task in prescription optimization. We integrate clustering with ranking in a heterogeneous information network.ResultsThe results from experiments on the Pharmacopoeia of the People’s Republic of China (ChP) demonstrate the effectiveness and accuracy of the proposed model for discovering useful categorizations of formulas.ConclusionsWe use heterogeneous information networks to model TCM data and propose a TCM-HIN. Combining the heterogeneous graph with the probability graph, we proposed the TCM-Clus algorithm, which combines clustering with ranking and classifies traditional Chinese medicine prescriptions. The results of the categorizations can help Chinese medicine practitioners to make clinical decision.

Graphic and Numerical Evolution of a Bonus-Malus System via Markov Chain Models

Abstract This paper describes an online solution for visualizing the simulation of a discrete time Markov chain using an animated graph. Based on the D3.js library, the development of this solution offers the user the choice between a presentation based on the concept of a probabilistic graph oriented and a presentation of the transitions in a matrix framework. As an application the solution helps to describe the dynamics of individual policyholders in a Bonus-Malus system of automobile insurance.

Incremental Pose Map Optimization for Monocular Vision SLAM Based on Similarity Transformation.

The novel contribution of this paper is to propose an incremental pose map optimization for monocular vision simultaneous localization and mapping (SLAM) based on similarity transformation, which can effectively solve the scale drift problem of SLAM for monocular vision and eliminate the cumulative error by global optimization. With the method of mixed inverse depth estimation based on a probability graph, the problem of the uncertainty of depth estimation is effectively solved and the robustness of depth estimation is improved. Firstly, this paper proposes a method combining the sparse direct method based on histogram equalization and the feature point method for front-end processing, and the mixed inverse depth estimation method based on a probability graph is used to estimate the depth information. Then, a bag-of-words model based on the mean initialization K-means is proposed for closed-loop feature detection. Finally, the incremental pose map optimization method based on similarity transformation is proposed to process the back end to optimize the pose and depth information of the camera. When the closed loop is detected, global optimization is carried out to effectively eliminate the cumulative error of the system. In this paper, indoor and outdoor environmental experiments are carried out using open data sets, such as TUM and KITTI, which fully proves the effectiveness of this method. Closed-loop detection experiments using hand-held cameras verify the importance of closed-loop detection. This method can effectively solve the scale drift problem of monocular vision SLAM and has strong robustness.

A FRAMEWORK FOR ONTOLOGY- BASED DIABETES DIAGNOSIS USING BAYELSIAN OPTIMIZATION TECHNIQUE

Diabetes Management System (DMS) is a computer-based system which aid physicians in properly diagnosing diabetes mellitus disease in patients. The DMS is essential in making individuals who have diabetes aware of their state and type. Existing approaches employed have not been efficient in considering all the diabetes type as well as making full prescription to diabetes patients. In this paper, a framework for an improved Ontology-based Diabetes Management System with a Bayesian optimization technique is presented. This helped in managing the diagnosis of diabetes and the prescription of treatment and drug to patients using the ontology knowledge management. The framework was implemented using Java programming language on Netbeans IDE, Protégé 4.2 and mysql. An extract of the ontology graph and acyclic probability graph was shown. The result showed that the nature of Bayesian network which has to do with statistical calculations based on equations, functions and sample frequencies led to more precise and reliable outcome.
 

An eigendecomposition method based on deep learning and probabilistic graph model

With the rapid development of computer, computer vision derived from computer vision has also made important progress in the field of image research. The extraction of image information is the most basic work in the field of image research. However, in the current environment, there is still a lack of effective methods to understand more complex image problems, such as image shape, material and illumination distribution in the environment. Eigenimage decomposition can be achieved by obtaining albedo eigenvalues and luminance eigenvalues. The color and illumination information of the image can be obtained more intuitively. Based on this, this paper proposes an intrinsic image decomposition method based on depth learning and probability graph model, in order to extract image information more accurately. Firstly, a deep convolution neural network is trained to decompose reflectivity image and shadow image. Then the conditional random field is used to optimize the reflectivity image and shadow image. The convolutional neural network designed in this paper obtains preliminary results through multi-scale architecture, deep supervision, step-by-step refinement of synthetic images and multi-stage training, which has been significantly improved compared with previous algorithms. Then the essential image and the corresponding gradient image are further optimized by conditional random field, and the eigenvalue image with richer details and clearer boundary can be obtained.

Tightly Coupled GNSS/INS Integration via Factor Graph and Aided by Fish-Eye Camera

GNSS/INS integrated solution has been extensively studied over the past decades. However, its performance relies heavily on environmental conditions and sensor costs. The GNSS positioning can obtain satisfactory performance in the open area. Unfortunately, its accuracy can be severely degraded in a highly urbanized area, due to the notorious multipath effects and none-line-of-sight (NLOS) receptions. As a result, excessive GNSS outliers occur, which causes a huge error in GNSS/INS integration. This paper proposes to apply a fish-eye camera to capture the sky view image to further classify the NLOS and line-of-sight (LOS) measurements. In addition, the raw INS and GNSS measurements are tightly integrated using a state-of-the-art probabilistic factor graph model. Instead of excluding the NLOS receptions, this paper makes use of both the NLOS and LOS measurements by treating them with different weightings. Experiments conducted in typical urban canyons of Hong Kong showed that the proposed method could effectively mitigate the effects of GNSS outliers, and an improved accuracy of GNSS/INS integration was obtained, when compared with the conventional GNSS/INS integration.

Fréchet analysis of variance for random objects

Summary Fréchet mean and variance provide a way of obtaining a mean and variance for metric space-valued random variables, and can be used for statistical analysis of data objects that lie in abstract spaces devoid of algebraic structure and operations. Examples of such data objects include covariance matrices, graph Laplacians of networks and univariate probability distribution functions. We derive a central limit theorem for the Fréchet variance under mild regularity conditions, using empirical process theory, and also provide a consistent estimator of the asymptotic variance. These results lead to a test for comparing $k$ populations of metric space-valued data objects in terms of Fréchet means and variances. We examine the finite-sample performance of this novel inference procedure through simulation studies on several special cases that include probability distributions and graph Laplacians, leading to a test for comparing populations of networks. The proposed approach has good finite-sample performance in simulations for different kinds of random objects. We illustrate the proposed methods by analysing data on mortality profiles of various countries and resting-state functional magnetic resonance imaging data.

Cover-up: a probabilistic privacy-preserving graph database model

A new graph database model is introduced that allows for an efficient and straightforward privacy-preserving mechanism. A probabilistic graph database model is also proposed, perhaps less suitable for lossless storage, but adapted for the use of statistical analysis that preserves the privacy of the individuals behind the data. Parallels are drawn to concepts in combinatorics such as clique complexes and incidence geometries.

Review, Evaluation, and Directions for Gene-Targeted Assembly for Ecological Analyses of Metagenomes.

Shotgun metagenomics has greatly advanced our understanding of microbial communities over the last decade. Metagenomic analyses often include assembly and genome binning, computationally daunting tasks especially for big data from complex environments such as soil and sediments. In many studies, however, only a subset of genes and pathways involved in specific functions are of interest; thus, it is not necessary to attempt global assembly. In addition, methods that target genes can be computationally more efficient and produce more accurate assembly by leveraging rich databases, especially for those genes that are of broad interest such as those involved in biogeochemical cycles, biodegradation, and antibiotic resistance or used as phylogenetic markers. Here, we review six gene-targeted assemblers with unique algorithms for extracting and/or assembling targeted genes: Xander, MegaGTA, SAT-Assembler, HMM-GRASPx, GenSeed-HMM, and MEGAN. We tested these tools using two datasets with known genomes, a synthetic community of artificial reads derived from the genomes of 17 bacteria, shotgun sequence data from a mock community with 48 bacteria and 16 archaea genomes, and a large soil shotgun metagenomic dataset. We compared assemblies of a universal single copy gene (rplB) and two N cycle genes (nifH and nirK). We measured their computational efficiency, sensitivity, specificity, and chimera rate and found Xander and MegaGTA, which both use a probabilistic graph structure to model the genes, have the best overall performance with all three datasets, although MEGAN, a reference matching assembler, had better sensitivity with synthetic and mock community members chosen from its reference collection. Also, Xander and MegaGTA are the only tools that include post-assembly scripts tuned for common molecular ecology and diversity analyses. Additionally, we provide a mathematical model for estimating the probability of assembling targeted genes in a metagenome for estimating required sequencing depth.

Применение комплексного подхода к определению остаточного ресурса промышленного здания с деформированными стропильными фермами

The article is devoted to the problem of objective estimation of residual resource based on a complex approach using probabilistic methods and graph theory. A single-storey building of the industrial workshop in Taganrog was chosen as an object for evaluation. The survey revealed typical defects and damages for industrial buildings - soaking, weathering of the solution, corrosion and absence of paint coating, as well as atypical defects - cuts of mechanical nature in various elements of farms. These damages are stress concentrators, so a three-dimensional model of the sling truss with damage simulation was created to calculate the stress-strain state at such sites and a finite element model was calculated in the ANSYS. According to the calculation results, it was found that in the most loaded section with defect (reference brace) there are significant stresses close to the design resistance of the soil, which reduces the total probability of failure-free operation of the truss. The calculation of the probability of failure-free operation for the truss was carried out using probabilistic methods, the relative reliability of the rest of the structures was determined on the basis of external characteristics and inspection of the characteristics of the structures by non-destructive testing methods. It has been established that the residual life in the calculation by the combined method is 17.3 years.

A review of Gaussian Markov models for conditional independence

Markov models lie at the interface between statistical independence in a probability distribution and graph separation properties. We review model selection and estimation in directed and undirected Markov models with Gaussian parametrization, emphasizing the main similarities and differences. These two model classes are similar but not equivalent, although they share a common intersection. We present the existing results from a historical perspective, taking into account the amount of literature existing from both the artificial intelligence and statistics research communities, where these models were originated. We cover classical topics such as maximum likelihood estimation and model selection via hypothesis testing, but also more modern approaches like regularization and Bayesian methods. We also discuss how the Markov models reviewed fit in the rich hierarchy of other, higher level Markov model classes. Finally, we close the paper overviewing relaxations of the Gaussian assumption and pointing out the main areas of application where these Markov models are nowadays used.

Tracking Tagged Inventory in Unstructured Environments through Probabilistic Dependency Graphs

Logging and tracking raw materials, workpieces and engineered products for seamless and quick pulls is a complex task in the construction and shipbuilding industries due to lack of structured storage solutions. Additional uncertainty is introduced if workpieces are stacked and moved by multiple stakeholders without maintaining an active and up-to-date log of such movements. While there are frameworks proposed to improve workpiece pull times using a variety of tracking modes based on deterministic approaches, there is little discussion of cases wherein direct observations are sparse due to occlusions from stacking and interferences. Our work addresses this problem by: logging visible part locations and timestamps, through a network of custom designed observation devices; and building a graph-based model to identify events that highlight part interactions and estimate stack formation to search for parts that are not directly observable. By augmenting the site workers and equipment with our wearable devices, we avoid adding additional cognitive effort for the workers. Native building blocks of the graph-based model were evaluated through simulations. Experiments were also conducted in an active shipyard to validate our proposed system.

White Learning: A White-Box Data Fusion Machine Learning Framework for Extreme and Fast Automated Cancer Diagnosis

Deep learning as a data modeling tool is hard to be understood of how its predicted result came about from its inner working. It is generally known as a black box and is not interpretable. Often in medical applications, physicians need to understand why a model predicts a result. On the other hand, BN is a probabilistic graph with nodes representing the variables, and the arcs present the conditional dependences between the variables. In this article, a white learning framework is proposed, which advocates three levels of fusing the black-box deep learning and white box BN, that offers both predictive power and interpretability. A case of breast cancer classification is conducted in an experiment. From the results, it is observed that white learning, which combines black-box and white-box machine learning, has an edge in performance over individually BN alone or deep learning alone. The white learning framework has the benefits of interpretability and high predictive power, making it suitable for critical decision-making task where a reliable prediction is as important as knowing how the outcome is predicted. The predicted output, which is generated from white learning, can be traced back via the conditional probability at each node. It is, hence, anticipated that in the future, especially for medical domain, white learning, which has the benefits of both black -box and white-box learning, would be highly valued and raised in popularity.

Research on statistical machine translation model based on deep neural network

With the increase of translation demand, the advancement of information technology, the development of linguistic theories and the progress of natural language understanding models in artificial intelligence research, machine translation has gradually gained worldwide attention. However, at present, machine translation research still has problems such as insufficient bilingual data and lack of effective feature representation, which affects the further improvement of key modules of machine translation such as word alignment, sequence adjustment and translation modelling. The effect of machine translation is still unsatisfactory. As a new machine learning method, deep neural network can automatically learn abstract feature representation and establish a complex mapping relationship between input and output signals, which provides a new idea for statistical machine translation research. Firstly, the multi-layer neural network and the undirected probability graph model are combined, and the similarity and context information of vocabulary are effectively utilized to model the word alignment more fully, and the word alignment model named NNWAM is constructed. Secondly, the low dimension will be used. The feature representation is combined with other features into a linearly ordered pre-ordering model to construct the pre-ordering model named NNPR. Finally, the word alignment model and the pre-ordering model are combined in the same deep neural network framework to form DNNAPM, a statistical machine translation model based on deep neural networks. The experimental results show that the statistical machine translation model based on deep neural network has better effect, faster convergence and better reliability than the comparison model algorithm.

Opinion spam detection by incorporating multimodal embedded representation into a probabilistic review graph

Spam reviews typically appear perfectly normal until examined in a large context. The standard approach to classifying reviews independently ignores these relations. In this study, we propose a complex probabilistic graph classification approach to address the problem of opinion spam detection. To obtain an initial effective spamicity estimation for the nodes (reviews, authors, and products) in the graph, we first train a neural network with attention mechanism to learn the multimodal embedded representation of nodes by leveraging both textual and rich features. Then based on the node prior computation, a heterogeneous graph is constructed to capture the relationships among different kinds of nodes, and the beliefs are further updated through iterative message propagation. To support this work, we collect two kinds of real-life datasets, which are separately composed of 97,839 restaurant reviews and 31,317 hotel reviews. The evaluation of the two datasets demonstrates the effectiveness of the proposed approach. We further analyze several salient rich features and the intermediate component of our model, thereby revealing that their states capture certain statistical characteristics of the datasets.

A Hybrid Markov Random Field Model With Multi-Granularity Information for Semantic Segmentation of Remote Sensing Imagery

High-spatial-resolution (HSR) remote sensing images usually contain rich hierarchical semantic information. However, many methods fail to solve the segmentation misclassification problems for HSR images due to just considering one layer of granularity information, such as the pixel granularity layer or the object granularity layer. This paper presents a hybrid Markov random field model for the semantic segmentation of HSR images by paying closer attention to capture the multi-granularity information. In this model, a probability graph with a multilayer structure is first built to represent different granularities information. Then, a hybrid label field is developed to model the multi-granularity classes with the form of a vector, and a new joint distribution is designed to capture the isotropic spatial interactions within each layer of the hybrid label field and the anisotropic spatial interactions between different layers. A generative probabilistic inference is proposed to realize the synergy between the multi-granularity information and the hybrid label interactions by iteratively updating the likelihood function and the joint probability of the hybrid label. The final semantic segmentation result can be achieved when the probabilistic inference converges. Experimental results over different HSR remote sensing images show that the proposed method can achieve more accurate segmentation than other state-of-the-art methods.

On the modeling of passenger mobility for stochastic bi-modal urban corridors

This paper proposes a methodology to estimate the passenger macroscopic fundamental diagram for bi-modal urban corridors while accounting for the stochastic nature of bus operations. The proposed framework extends the existing variational theory (VT) approaches by: (i) introducing a probabilistic VT graph, where the costs are computed using an efficient stochastic shortest path algorithm; (ii) capturing the effects of stochastic moving bus bottlenecks and the correlation of bus arrival times; (iii) incorporating a macroscopic passenger model that reflects the passenger dynamics for the different modes; and (iv) accounting for the effects that the traffic conditions might have on bus operations.Using a Monte-Carlo simulation and empirical data from a bi-modal corridor in Zurich, Switzerland, we not only successfully validate the results yielded by our stochastic VT approach, but also show its applicability on a real corridor. A comparison with a deterministic VT approach reveals the value of the proposed framework, especially for corridors with a high bus frequency and considerable stochasticity. The results demonstrate that incorporating stochasticity and the traffic conditions is essential if buses run with relatively short and variable headways. Moreover, we introduce an innovative application example for the evaluation of different bus lane layouts, aiming to maximize the passenger throughput along a bi-modal urban corridor. The application example shows that the proposed framework can be used as an efficient modeling tool for practitioners. In particular, it can be used to identify a proper lane allocation strategy by computing the critical density of cars when a mixed lane should be switched to a dedicated bus lane or vice versa. It is important to note that such application would not have been possible without our proposed VT extensions, which account for both passenger dynamics and the impact of traffic conditions. Finally, given that the proposed methodology is generic, it can be easily extended to various traffic problems involving stochasticity.