Search

Although deep learning for Diabetic Retinopathy (DR) screening has shown great success in achieving clinically acceptable accuracy for referable versus non-referable DR, there remains a need to provide more fine-grained grading of the DR severity level as well as automated segmentation of lesions (if any) in the retina images. We observe that the DR severity level of an image is dependent on the presence of different types of lesions and their prevalence. In this work, we adopt a multi-task learning approach to perform the DR grading and lesion segmentation tasks. In light of the lack of lesion segmentation mask ground-truths, we further propose a semi-supervised learning process to obtain the segmentation masks for the various datasets. Experiments results on publicly available datasets and a real world dataset obtained from population screening demonstrate the effectiveness of the multi-task solution over state-of-the-art networks.

Deep learning-based models are developed to automatically detect if a retina image is `referable' in diabetic retinopathy (DR) screening. However, their classification accuracy degrades as the input images distributionally shift from their training distribution. Further, even if the input is not a retina image, a standard DR classifier produces a high confident prediction that the image is `referable'. Our paper presents a Dirichlet Prior Network-based framework to address this issue. It utilizes an out-of-distribution (OOD) detector model and a DR classification model to improve generalizability by identifying OOD images. Experiments on real-world datasets indicate that the proposed framework can eliminate the unknown non-retina images and identify the distributionally shifted retina images for human intervention., Accepted at IEEE ICIP 2021

Patient similarity analytics has emerged as an essential tool to identify cohorts of patients who have similar clinical characteristics to some specific patient of interest. In this study, we propose a patient similarity measure called D3K that incorporates domain knowledge and data-driven insights. Using the electronic health records (EHRs) of 169,434 patients with either diabetes, hypertension or dyslipidaemia (DHL), we construct patient feature vectors containing demographics, vital signs, laboratory test results, and prescribed medications. We discretize the variables of interest into various bins based on domain knowledge and make the patient similarity computation to be aligned with clinical guidelines. Key findings from this study are: (1) D3K outperforms baseline approaches in all seven sub-cohorts, (2) our domain knowledge-based binning strategy outperformed the traditional percentile-based binning in all seven sub-cohorts, (3) there is substantial agreement between D3K and physicians (κ = 0.746), indicating that D3K can be applied to facilitate shared decision making. This is the first study to use patient similarity analytics on a cardiometabolic syndrome-related dataset sourced from medical institutions in Singapore. We consider patient similarity among patient cohorts with the same medical conditions to develop localized models for personalized decision support to improve the outcomes of a target patient.

Microblogs have become the preferred means of communication for people to share information and feelings, especially for fast evolving events. Understanding the emotional reactions of people allows decision makers to formulate policies that are likely to be more well-received by the public and hence better accepted especially during policy implementation. However, uncovering the topics and emotions related to an event over time is a challenge due to the short and noisy nature of microblogs. This work proposes a weakly supervised learning approach to learn coherent topics and the corresponding emotional reactions as an event unfolds. We summarize the event by giving the representative microblogs and the emotion distributions associated with the topics over time. Experiments on multiple real-world event datasets demonstrate the effectiveness of the proposed approach over existing solutions.

Link prediction is a fundamental challenge in network science. Among various methods, local similarity indices are widely used for their high cost-performance. However, the performance is less robust: for some networks local indices are highly competitive to state-of-the-art algorithms while for some other networks they are very poor. Inspired by techniques developed for recommender systems, we propose an enhancement framework for local indices based on collaborative filtering (CF). Considering the delicate but important difference between personalized recommendation and link prediction, we further propose an improved framework named as self-included collaborative filtering (SCF). The SCF framework significantly improved the accuracy and robustness of well-known local indices. The combination of SCF framework and a simple local index can produce an index with competitive performance and much lower complexity compared with elaborately-designed state-of-the-art algorithms., 19 pages, 3 tables, 1 figure

Though deep learning systems have achieved high accuracy in detecting diseases from medical images, few such systems have been deployed in highly automated disease screening settings due to lack of trust in how well these systems can generalize to out-of-datasets. We propose to use uncertainty estimates of the deep learning system’s prediction to know when to accept or to disregard its prediction. We evaluate the effectiveness of using such estimates in a real-life application for the screening of diabetic retinopathy. We also generate visual explanation of the deep learning system to convey the pixels in the image that influences its decision. Together, these reveal the deep learning system’s competency and limits to the human, and in turn the human can know when to trust the deep learning system.

Ischemic stroke is a leading cause of death and long-term disability that is difficult to predict reliably. Retinal fundus photography has been proposed for stroke risk assessment, due to its non-invasiveness and the similarity between retinal and cerebral microcirculations, with past studies claiming a correlation between venular caliber and stroke risk. However, it may be that other retinal features are more appropriate. In this paper, extensive experiments with deep learning on six retinal datasets are described. Feature isolation involving segmented vascular tree images is applied to establish the effectiveness of vessel caliber and shape alone for stroke classification, and dataset ablation is applied to investigate model generalizability on unseen sources. The results suggest that vessel caliber and shape could be indicative of ischemic stroke, and sourcespecific features could influence model performance.

Link prediction is a fundamental challenge in network science. Among various methods, similarity-based algorithms are popular for their simplicity, interpretability, high efficiency and good performance. In this paper, we show that the most elementary local similarity index Common Neighbor (CN) can be linearly decomposed by eigenvectors of the adjacency matrix of the target network, with each eigenvector's contribution being proportional to the square of the corresponding eigenvalue. As in many real networks, there is a huge gap between the largest eigenvalue and the second largest eigenvalue, the CN index is thus dominated by the leading eigenvector and much useful information contained in other eigenvectors may be overlooked. Accordingly, we propose a parameter-free algorithm that ensures the contributions of the leading eigenvector and the secondary eigenvector the same. Extensive experiments on real networks demonstrate that the prediction performance of the proposed algorithm is remarkably better than well-performed local similarity indices in the literature. A further proposed algorithm that can adjust the contribution of leading eigenvector shows the superiority over state-of-the-art algorithms with tunable parameters for its competitive accuracy and lower computational complexity., Comment: 22 pages, 4 figures, 4 tables

Learning concepts that are consistent with human perception is important for Deep Neural Networks to win end-user trust. Post-hoc interpretation methods lack transparency in the feature representations learned by the models. This work proposes a guided learning approach with an additional concept layer in a CNN- based architecture to learn the associations between visual features and word phrases. We design an objective function that optimizes both prediction accuracy and semantics of the learned feature representations. Experiment results demonstrate that the proposed model can learn concepts that are consistent with human perception and their corresponding contributions to the model decision without compromising accuracy. Further, these learned concepts are transferable to new classes of objects that have similar concepts., Comment: Accepted to IJCNN 2021

Automated fact-checking on a large-scale is a challenging task that has not been studied systematically until recently. Large noisy document collections like the web or news articles make the task more difficult. We describe a three-stage automated fact-checking system, named Quin+, using evidence retrieval and selection methods. We demonstrate that using dense passage representations leads to much higher evidence recall in a noisy setting. We also propose two sentence selection approaches, an embedding-based selection using a dense retrieval model, and a sequence labeling approach for context-aware selection. Quin+ is able to verify open-domain claims using results from web search engines.

Smart city and smart hospital development comprise a mainstream strategy worldwide to enhance the achievement of Sustainable Development Goals (SDGs) and Corporate Social Responsibility (CSR) practices. The smart healthcare industry needs to optimize its smart healthcare information strategies, creating the process-activity value and reducing the cost of healthcare without sacrificing the quality of patient care. This study proposes an decision model of Activity-Based Costing (ABC) concept and Multi-Criteria Decision-Making (MCDM) techniques that integrates the Decision-Making Trial and Evaluation Laboratory (DEMATEL), Analytic Network Processes (ANP), the VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) and Zero-One Goal Programming (ZOGP) to achieve optimal smart healthcare information system portfolio strategy decisions. The resulting data shows that the perspective of government policy support and the criterion of Minimize Equipment Maintenance Costs are the most significant evaluation factors, and that the Health Data Informational System and Hospital Device and Drug Management System comprise by far the optimal portfolio of smart healthcare information system which strengthen the connection between ABC and MCDM techniques in evaluation process. The major contributions of this study are as follows: (1) the proposed model contributes to the management accounting innovation development of a sustainable city and Operation Research (OR) application, (2) the integrated model can promote the smart healthcare industry development and help decision makers to more accurately understand how to allocate resources and planning for intelligent-related activities to each smart healthcare information system through appropriate cost drivers.

Bayesian approaches have been successfully applied in social network analysis to study group behaviors such as online information dissemination and voting pattern. The focus has been on estimating the structure and strength of peer influence and its impact on the decisions of an individual. Less attention has been given to incorporating contextual information and individuals' hidden characteristics (or bias). In this work, we examine the social dynamics where social influence and contextual information play pivotal roles in driving one's decision. We design a probabilistic graphical model called CLAP to understand users' decision behavior in a social network, with an emphasis on both social-level and individual-level factors. To this end, the proposed model introduces hidden bias states associated with each actor and jointly estimates each actor's hidden bias state together with the social influence network. We demonstrate the effectiveness of CLAP on two types of social networks, a real-world US Congress network where senators vote on new bills, and online Twitter networks where users debate on the effectiveness of vaccine and lockdown policy during COVID-19. The experiment results show that CLAP outperforms state-of-the-art game theoretic approaches in predicting user decision. Further, the estimated social influence networks by CLAP has high edge homogeniety ratios.

Existing defenses against adversarial attacks are typically tailored to a specific perturbation type. Using adversarial training to defend against multiple types of perturbation requires expensive adversarial examples from different perturbation types at each training step. In contrast, manifold-based defense incorporates a generative network to project an input sample onto the clean data manifold. This approach eliminates the need to generate expensive adversarial examples while achieving robustness against multiple perturbation types. However, the success of this approach relies on whether the generative network can capture the complete clean data manifold, which remains an open problem for complex input domain. In this work, we devise an approximate manifold defense mechanism, called RBF-CNN, for image classification. Instead of capturing the complete data manifold, we use an RBF layer to learn the density of small image patches. RBF-CNN also utilizes a reconstruction layer that mitigates any minor adversarial perturbations. Further, incorporating our proposed reconstruction process for training improves the adversarial robustness of our RBF-CNN models. Experiment results on MNIST and CIFAR-10 datasets indicate that RBF-CNN offers robustness for multiple perturbations without the need for expensive adversarial training., Workshop on Machine Learning with Guarantees, NeurIPS 2019. IJCNN, 2020 (full paper)

The purpose of this paper is to create a smart operating roadmap, which shows the entire process of a strategic business plan, including functions, methods, and tools, to link IFRS 8 (International Financial Reporting Standards No.8) to ABSC (Activity-Based Standard Costing), and to integrate ERP (Enterprise Resource Planning), MES (Manufacturing Execution System) under an Industry 4.0 environment. The IFRS is a global accounting framework that provides high-quality global accounting standards and governance principles for companies. Using the ABSC production decision model can support the Chief Operating Decision Maker (CODM) in planning Product&ndash, Business Unit (Product-BU) organization, which complies with the definition of the IFRS8 operating segments. The case study of the steel group uses the organizational design of the ERP system to achieve the systematization of reportable segments financial statements. In this process, the mathematical programing methods and ABSC can be used to obtain the optimal solutions for sales, costs, and profits. An international steel group case is used to demonstrate how to apply the methodology proposed in this paper for operating planning and control. The sensitivity analysis on the carbon emission reduction goal of environmental sustainability is also presented for the steel group case.

To harness the rich amount of information available on the web today, many organizations aggregate public (and private) data to derive knowledge repositories for real-world entities. This paper aims to build historical profiles of real-world entities by integrating temporal records collected from different sources. This problem is challenging not only because entities may change their attribute values over time, but also because information provided by the sources could be unreliable. In this paper, we present a new solution for profiling entities over time. To understand the evolution of entities, we describe a novel transition model which gives the probability that an entity will change to a particular attribute value after some time period. Next, a set of quality metrics are defined for the data sources to capture the exactness and timeliness of their provided values. The transition model and the quality metrics are then built into a source-aware temporal matching algorithm that can link temporal records to entities at the right time and augment entity profiles with correct values. Our suite of experiments demonstrate that the proposed approach is able to outperform the state-of-the-art techniques by constructing more complete and accurate profiles for entities.

Link prediction is a significant and challenging task in network science. The majority of known methods are similarity-based, which assign similarity indices for node pairs and assume that two nodes of larger similarity have higher probability to be connected by a link. Due to their simplicity, interpretability and high efficiency, similarity-based methods, in particular those based only on local information, have already found successful applications on disparate fields. In this research domain, an intuitive consensus is that two nodes sharing common neighbors are very likely to have a link, while some recent evidences argue that the number of 3-hop paths more accurately predicts missing links than the number of common neighbors. In this paper, we implement extensive experimental comparisons between 2-hop-based and 3-hop-based similarity indices on 128 real networks. Our results indicate that the 3-hop-based indices perform slightly better with a winning rate about 55.88%, but which index is the best one still depends on the target network. Overall speaking, the class of Cannistraci-Hebb indices performs the best among all considered candidates., 4 figures

Link prediction is an elemental challenge in network science, which has already found applications in guiding laboratorial experiments, digging out drug targets, recommending friends in social networks, probing mechanisms in network evolution, and so on. With a simple assumption that the likelihood of the existence of a link between two nodes can be unfolded by a linear summation of neighboring nodes' contributions, we obtain the analytical solution of the optimal likelihood matrix, which shows remarkably better performance in predicting missing links than the state-of-the-art algorithms for not only simple networks, but also weighted and directed networks. To our surprise, even some degenerated local similarity indices from the solution outperform well-known local indices, which largely refines our knowledge, for example, the direct count of the number of 3-hop paths between two nodes more accurately predicts missing links than the number of 2-hop paths (i.e., the number of common neighbors), while in the previous studies, as indicated by the local path index and Katz index, the statistics on longer paths are always considered to be complementary to but less important than those on shorter paths., 29 pages, 3 figures and 5 tables

User opinions expressed in the form of ratings can influence an individual's view of an item. However, the true quality of an item is often obfuscated by user biases, and it is not obvious from the observed ratings the importance different users place on different aspects of an item. We propose a probabilistic modeling of the observed aspect ratings to infer (i) each user's aspect bias and (ii) latent intrinsic quality of an item. We model multi-aspect ratings as ordered discrete data and encode the dependency between different aspects by using a latent Gaussian structure. We handle the Gaussian-Categorical non-conjugacy using a stick-breaking formulation coupled with P\'{o}lya-Gamma auxiliary variable augmentation for a simple, fully Bayesian inference. On two real world datasets, we demonstrate the predictive ability of our model and its effectiveness in learning explainable user biases to provide insights towards a more reliable product quality estimation., Comment: Accepted for publication in IJCAI 2017

User generated content about products and services in the form of reviews are often diverse and even contradictory. This makes it difficult for users to know if an opinion in a review is prevalent or biased. We study the problem of searching for supporting opinions in the context of reviews. We propose a framework called SURF, that first identifies opinions expressed in a review, and then finds similar opinions from other reviews. We design a novel probabilistic graphical model that captures opinions as a combination of aspect, topic and sentiment dimensions, takes into account the preferences of individual authors, as well as the quality of the entity under review, and encodes the flow of thoughts in a review by constraining the aspect distribution dynamically among successive review segments. We derive a similarity measure that considers both lexical and semantic similarity to find supporting opinions. Experiments on TripAdvisor hotel reviews and Yelp restaurant reviews show that our model outperforms existing methods for modeling opinions, and the proposed framework is effective in finding supporting opinions.

Many applications require understanding how event occurrences at one geographical region affect or influence event occurrences at another region, e.g. spread of disease and forest fires. Existing works typically impose a grid to partition the spatial space and utilize spatial autocorrelation property to model the spatial dependency among the grid cells. However, they are often highly sensitive to the granularity of the grid size and they do not incorporate the temporal dynamics of the event occurrences among regions. This paper utilizes the notion of a spatial network with temporal dependency to capture the dynamics of event occurrences among regions. This network is modeled as a directed graph where each node is a group of spatially nearby events and each directed edge represents the influence of events from a source node to a destination node. We design an algorithm called LinkNet to generate this network from spatio-temporal event databases. LinkNet utilizes minimum description length based information---theoretic approach to automatically adjust the number of regions and the temporal relationships among regions. Two optimizations are devised to reduce the computational complexity of LinkNet. We also demonstrate how the proposed network can be used for hotspot prediction. Experiment results on both synthetic and real world datasets demonstrate the efficiency of LinkNet and the effectiveness of the network in predicting the next hotspots.

Identifying influential nodes in networks is a significant and challenging task. Among many centrality indices, the $k$-shell index performs very well in finding out influential spreaders. However, the traditional method for calculating the $k$-shell indices of nodes needs the global topological information, which limits its applications in large-scale dynamically growing networks. Recently, L\@\"{u} \emph{et al.} [Nature Communications 7 (2016) 10168] proposed a novel asynchronous algorithm to calculate the $k$-shell indices, which is suitable to deal with large-scale growing networks. In this paper, we propose two algorithms to select nodes and update their intermediate values towards the $k$-shell indices, which can help in accelerating the convergence of the calculation of $k$-shell indices. The former algorithm takes into account the degrees of nodes while the latter algorithm prefers to choose the node whose neighbors' values have been changed recently. We test these two methods on four real networks and three artificial networks. The results suggest that the two algorithms can respectively reduce the convergence time up to 75.4\% and 92.9\% in average, compared with the original asynchronous updating algorithm., Comment: 18 pages, 4 figures, 2 tables

This paper examines genre-related tags through a social bookmarking system CiteULike. Three types of genre-related tags were identified: “is” tag indicating a genre,“for” tag for creating another genre, and “about” tag dealing with a genretopic. An analysis of the frequency and functions of genre-related tags will enable us better understand tagging as well as genre.

This study investigates the feasibility of an ensemble of classifiers in characterizing ultrasonic liver tissue. Texture analysis generally requires feature representation and classification algorithm. From a variety of feature representations and classification algorithms, obtaining optimal ensembles composed of any feature-classifier pairs is difficult. This paper proposes an ensemble creation algorithm that can form an ensemble with high generalization performance. The pattern recognition process comprises four main stages. The first stage utilized multiresolution analysis to extract intrinsic features of ultrasonic liver images. By utilizing spatial-frequency decomposition, a feature vector was obtained by collecting the feature representation for each subimage. In the second stage of the study, various classification algorithms with diverse feature vectors were trained. Based on the trained classifiers, an ensemble was created by using the proposed algorithm in the third stage. The last stage was concerned with the aggregation of individual classifiers. The proposed approach was applied to discriminate ultrasonic liver images from three liver states: normal liver, cirrhosis, and hepatoma. Based on the six well-known fusion schemes, the experimental results showed that the ensemble proposed in this study yields more discrimination. The results indicate that the combining multiple classifiers with different features is an effective approach for characterizing ultrasonic live r tissue. Furthermore, a clinician can use the quantitative index of the classification results when deciding whether to conduct an advanced medical examination, thus improving the quality of medical care.

Internet addiction and other such problems stem from the high frequency of Internet usage of adolescents. The study aims at working out the counseling strategy proposed to prevent internet addiction among the youth. In this study, DEMATEL was the avenue for creating a discussion with families, schools, the society and analyzing regulations & policies so as to find out the cause-and-effect relationship among them and work out the most efficient counseling strategy. The study result shows that the family plays the most crucial role in curing the adolescents’ Internet addiction through action.

Highlights? We extract five feature spaces for the ultrasonic liver tissue characterization. ? The proposed genetic-algorithm-based feature fusion is effective. ? Feature fusion combines features selected from different feature spaces. ? We select features to improve the classification accuracy for US liver images. This paper describes a two-stage feature fusion method for ultrasonic liver tissue characterization. The proposed method hierarchically incorporates a genetic-algorithm-based feature selection to automatically select more efficient feature subset to discriminate among ultrasonic images of liver tissue in three states: normal liver, cirrhosis, and hepatoma. Multiple feature spaces are adopted in this paper, including the spatial gray-level dependence matrices (SGLDMs), multiresolution fractal feature vector and multiresolution energy feature vector. Features extracted from different feature spaces may contain complementary information. The feature subsets of different feature spaces are fused and the genetic-algorithm-based feature selection is applied onto the fused feature space to facilitate the two-stage feature fusion. The classification accuracy of the fused feature subset is up to 96.62%. Experimental results demonstrate that the proposed method is capable to select discriminative features among multiple feature vectors to achieve the early detection of hepatoma and cirrhosis based on ultrasonic liver imaging.

The fundamental principle of order in the library catalogue is the authorship principle, which serves as the organizing node of an alphabetico-classed system, in which “texts” of “works” are organized first alphabetically by uniform title of the progenitor work and then are subarranged using titles for variant instantiations, under the heading for an “author.” We analyze case studies of entries from (1) the first documented imperial library catalogue, the Seven Epitomes (Qilue [七略]), in China; (2) Abelard’s Works, which featured prominently in the 1848 testimony of Antonio Panizzi; and (3)The French Chef and the large family of instantiated works associated with it. Our analysis shows that the catalogue typically contains many large superwork sets. A more pragmatic approach to the design of catalogues is to array descriptions of resources in relation to the superwork sets to which they might belong. In all cases, a multidimensional faceted arrangement incorporating ideational nodes from the universe of recorded knowledge holds promise for greatly enhanced retrieval capability.

When two-electrode spark gap switches are applied to main discharging circuit of pulsed power conditioning system, trigger generators will be needed. To guarantee triggering reliability and protect other key components in the pulsed power conditioning system, such as pre-ionization branch and xenon lamp, magnetic switch is needed to insulate the high triggering voltage in the main discharging circuit. In this paper, starting from the analysis of the working principle of the discharging circuit, we present the solution of Fe-based nanocrystalline alloy switch. Specifically, we provide the basis of theoretical calculation and the method of material selection for the design and realization of this magnetic switch. Via the performance tests in pulsed power conditioning system, the validity of material selection and parameter design is proved. This paper provides a reference for the engineering design and application of magnetic switches in huge pulsed power systems.

A recent approach to improve the performance of XML query evaluation is to cache the query results of frequent query patterns. Unfortunately, discovering these frequent query patterns is an expensive operation. In this paper, we develop a two-pass mining algorithm 2PXMiner that guarantees the discovery of frequent query patterns by scanning the database at most twice. By exploiting a transaction summary data structure, and an enumeration tree, we are able to determine the upper bounds of the frequencies of the candidate patterns, and to quickly prune away the infrequent patterns. We also design an index to trace the repeating candidate subtrees generated by sibling repetition, thus avoiding redundant computations. Experiments results indicate that 2PXMiner is both efficient and scalable.

Existing systems that support semistructured views do not maintain semantics during the process of designing the views. Thus, these systems do not guarantee the validity and reversibility of the views. In this paper, we propose an approach to address the issue of valid and reversible semistructured views. We design a set of view operators for designing semistructured views. These operators are select, drop, join and swap. For each operator, we develop a complete set of rules to maintain the semantics of the views. In particular, we maintain the evolution and integrity of relationships once an operator is applied. We also examine the reversible view problem under our operators and develop rules to guarantee that the designed views are reversible. Finally, we examine the changes in the participation constraints of relationship types during the view design process, and develop rules to ensure the correctness of the participation constraints.

The physical implementation of a prototypical 250-MHz CMOS 4-T SRAM is described in this paper. The proposed SRAM cell takes advantage of a negative word-line scheme to minimize the leakage current of the cell access transistors. As a result, the standby power consumption is drastically reduced. The proposed 4-kb 4-T SRAM is measured to consume 0.32 mW in the standby mode, and a 3.8-ns access time in the R/W mode. The highest operating clock rate is measured to be 263 MHz

This work addresses the problem of finding boundary points in multidimensional data sets. Boundary points are data points that are located at the margin of densely distributed data such as a cluster. We describe a novel approach called BORDER (a BOundaRy points DEtectoR) to detect such points. BORDER employs the state-of-the-art database technique - the Gorder kNN join and makes use of the special property of the reverse k nearest neighbor (RkNN). Experimental studies on data sets with varying characteristics indicate that BORDER is able to detect the boundary points effectively and efficiently.

Objective:: Current protein-protein interaction (PPI) detection via high-throughput experimental methods, such as yeast-two-hybrid has been reported to be highly erroneous, leading to potentially costly spurious discoveries. This work introduces a novel measure called IRAP, i.e. ''interaction reliability by alternative path'', for assessing the reliability of protein interactions based on the underlying topology of the PPI network. Methods and materials:: A candidate PPI is considered to be reliable if it is involved in a closed loop in which the alternative path of interactions between the two interacting proteins is strong. We devise an algorithm called AlternativePathFinder to compute the IRAP value for each interaction in a complex PPI network. Validation of the IRAP as a measure for assessing the reliability of PPIs is performed with extensive experiments on yeast PPI data. All the data used in our experiments can be downloaded from our supplementary data web site at http://www.comp.nus.edu.sg/~chenjin/data.html. Results:: Results show consistently that IRAP measure is an effective way for discovering reliable PPIs in large datasets of error-prone experimentally-derived PPIs. Results also indicate that IRAP is better than IG2, and markedly better than the more simplistic IG1 measure. Conclusion:: Experimental results demonstrate that a global, system-wide approach-such as IRAP that considers the entire interaction network instead of merely local neighbors-is a much more promising approach for assessing the reliability of PPIs.

New clinical studies suggest that narrowing of the retinal blood vessels may be an early indicator of cardiovascular diseases. One measure to quantify the severity of retinal arteriolar narrowing is the arteriolar-to-venular diameter ratio (AVR). The manual computation of AVR is a tedious process involving repeated measurements of the diameters of all arterioles and venules in the retinal images by human graders. Consistency and reproducibility are concerns. To facilitate large-scale clinical use in the general population, it is essential to have a precise, efficient and automatic system to compute this AVR. This paper describes a new approach to obtain AVR. The starting points of vessels are detected using a matched Gaussian filter. The detected vessels are traced with the help of a combined Kalman filter and Gaussian filter. A modified Gaussian model that takes into account the central light reflection of arterioles is proposed to describe the vessel profile. The width of a vessel is obtained by data fitting. Experimental results indicate a 97.1% success rate in the identification of vessel starting points, and a 99.2% success rate in the tracking of retinal vessels. The accuracy of the AVR computation is well within the acceptable range of deviation among the human graders, with a mean relative AVR error of 4.4%. The system has interested clinical research groups worldwide and will be tested in clinical studies.

A low-cost video decoder for NTSC signals is described in this paper. The proposed NTSC video decoder design employs a 2D2L comb filter, and a DDFS (direct digital frequency synthesizer)-based DCO (digital control oscillator) based on a trigonometric quadruple angle formula in a digital PLL to track and lock the demodulation clocks. The complexity of the digital video decoder is consequently drastically reduced. The overall cost of the proposed design is 7.22 mm/sup 2/ (26 K gates). The maximum power dissipation is 109.2 mW at the highest clock rate which is 21.48 MHz.

Efficient clustering in dynamic spatial databases is currently an open problem with many potential applications. Most traditional spatial clustering algorithms are inadequate because they do not have an efficient support for incremental clustering.In this paper, we propose DClust, a novel clustering technique for dynamic spatial databases. DClust is able to provide multi-resolution view of the clusters, generate arbitrary shapes clusters in the presence of noise, generate clusters that are insensitive to ordering of input data and support incremental clustering efficiently. DClust utilizes the density criterion that captures arbitrary cluster shapes and sizes to select a number of representative points, and builds the Minimum Spanning Tree (MST) of these representative points, called R-MST. After the initial clustering, a summary of the cluster structure is built. This summary enables quick localization of the effect of data updates on the current set of clusters. Our experimental results show that DClust outperforms existing spatial clustering methods such as DBSCAN, C2P, DENCLUE, Incremental DBSCAN and BIRCH in terms of clustering time and accuracy of clusters found.

Remote homology detection refers to the detection of structural homology in proteins when there is little or no sequence similarity. In this article, we present a remote homolog detection method called SVM-HMMSTR that overcomes the reliance on detectable sequence similarity by transforming the sequences into strings of hidden Markov states that represent local folding motif patterns. These state strings are transformed into fixed-dimension feature vectors for input to a support vector machine. Two sets of features are defined: an order-independent feature set that captures the amino acid and local structure composition; and an order-dependent feature set that captures the sequential ordering of the local structures. Tests using the Structural Classification of Proteins (SCOP) 1.53 data set show that the SVM-HMMSTR gives a significant improvement over several current methods.

Data quality problems can arise from abbreviations, data entry mistakes, duplicate records, missing fields, and many other sources. These problems proliferate when you integrate multiple data sources in data warehousing, federated databases, and global information systems. A newly discovered class of erroneous data is spurious links, where a real-world entity has multiple links that might not be properly associated with it. The existence of such spurious links often leads to confusion and misrepresentation in the data records representing the entity. Although the data set is well known for its high-quality bibliographic information, collecting and maintaining the data from diverse sources requires enormous effort. Errors, including spurious links, are inevitable. To solve this problem, we use context information to identify spurious links. First, we identify data records that contain potential spurious links. We then determine the set of attributes that constitute each record's context. Experiments with three real-world databases have demonstrated that our approach can accurately identify spurious links. Comparing context information between data records can help solve the data quality problem of spurious links-that is, multiple links between data entries and real-world entities.

This paper describes the feasibility of selecting fractal feature vector based on M-band wavelet transform to classify ultrasonic liver images-normal liver, cirrhosis, and hepatoma. The proposed feature extraction algorithm is based on the spatial-frequency decomposition and fractal geometry. Various classification algorithms based on respective texture measurements and filter banks are presented and tested. Classifications for the three sets of ultrasonic liver images reveal that the fractal feature vector based on M-band wavelet transform is trustworthy. A hierarchical classifier, which is based on the proposed feature extraction algorithm is at least 96.7% accurate in the distinction between normal and abnormal liver images and is at least 93.6% accurate in the distinction between cirrhosis and hepatoma liver images. Additionally, the criterion for feature selection is specified and employed for performance comparisons herein.

The Extensible Markup Language has become the standard for information interchange on the Web. We study the data- and document-centric uses of XML management systems (XMLMS). We want to provide XML data users with a guideline for choosing the data management system that best meets their needs. Because the systems we test are first-generation approaches, we suggest a hypothetical design for a useful XML database that could use all the expressive power of XML and XML query languages.

Concept lattice model, the core structure in formal concept analysis, has been successfully applied in software engineering and knowledge discovery. This paper integrates the simple base classifier (Naive Bayes or Nearest Neighbour) into each node of the concept lattice to form a new composite classifier. Two new classification systems are developed, CLNB and CLNN, which employ efficient constraints to search for interesting patterns and voting strategy to classify a new object. CLNB integrates the Naiive Bayes base classifier into concept nodes while CLNN incorporates the Nearest Neighbour base classifier into concept nodes. Experimental results indicate that these two composite classifiers greatly improve the accuracy of their corresponding base classifier. In addition, CLNB even outperforms three other state-of-the-art classification methods, NBTree, CBA and C4.5 Rules.

Advances in image acquisition and storage technology have led to tremendous growth in very large and detailed image databases. These images, if analyzed, can reveal useful information to the human users. Image mining deals with the extraction of implicit knowledge, image data relationship, or other patterns not explicitly stored in the images. Image mining is more than just an extension of data mining to image domain. It is an interdisciplinary endeavor that draws upon expertise in computer vision, image processing, image retrieval, data mining, machine learning, database, and artificial intelligence. In this paper, we will examine the research issues in image mining, current developments in image mining, particularly, image mining frameworks, state-of-the-art techniques and systems. We will also identify some future research directions for image mining.

Existing duplicate elimination methods for data cleaning work on the basis of computing the degree of similarity between nearby records in a sorted database. High recall can be achieved by accepting records with low degrees of similarity as duplicates, at the cost of lower precision. High precision can be achieved analogously at the cost of lower recall. This is the recall–precision dilemma . We develop a generic knowledge-based framework for effective data cleaning that can implement any existing data cleaning strategies and more. We propose a new method for computing transitive closure under uncertainty for dealing with the merging of groups of inexact duplicate records and explain why small changes to window sizes has little effect on the results of the sorted neighborhood method. Experimental study with two real-world datasets show that this approach can accurately identify duplicates and anomalies with high recall and precision, thus effectively resolving the recall–precision dilemma.

The advancement in strategies for bone and cartilage regeneration have been centralized on a concept that describes the close relationship between osteogenic cells, osteoconductive scaffolds, delivery growth factors and its mechanical environment. The dynamic nature of tissue repair process involves intricate mimicry of signals expressed in the biological system in response to an injury. Recently, synergistic strategies involving hybrid delivery systems that provide sequential dual delivery of biomolecules and relevant topological cues received great attention. Future advances in tissue regeneration will therefore depend on multidisciplinary strategies that encompass the crux of tissue repair aimed towards constructing the ideal functional regenerative scaffold. Here, functional scaffolds delivering therapeutics are reviewed in terms of their controlled release and healing capabilities. ASTAR (Agency for Sci., Tech. and Research, S’pore) Accepted version

This paper presents a new ocean platform concept, the fully constrained platform (FCP). In a nut-shell, FCP is a fixed-in-place deepwater platform concept, which has the least amount of motion in harsh environment, compared to all other deepwater platforms currently in use. For an intermediate water depth, from 100 to 1,000 ft, the FCP concept could be competitive to existing designs. FCP is the core technology for a new class of marine structures, and has the potential to be used anywhere in world’s ocean for any purpose, which requires a support structure with minimum motions. Paper published with permission.

Research in relational keyword search has been focused on the efficient computation of results as well as strategies to rank and output the most relevant ones. However, the challenge to retrieve the intended results remains. Existing relational keyword search techniques suffer from the problem of returning overwhelming number of results, many of which may not be useful. In this work, we adopt a semantic approach to relational keyword search via an Object-Relationship-Mixed data graph. This graph is constructed based on database schema constraints to capture the semantics of objects and relationships in the data. Each node in the ORM data graph represents either an object, or a relationship, or both. We design an algorithm that utilizes the ORM data graph to process keyword queries. Experiment results show our approach returns more informative results compared to existing methods, and is efficient.

The traditional problem of updating relational databases through views is an important practical problem that has attracted much interest. In this paper, we examine the problem of view update in Entity-Relationship based database management systems [1] where the conceptual schema is represented by a normal form ER diagram [2] and views may be modelled by ER diagrams. We develop a theory within the framework of the ER approach that characterizes the conditions under which there exist mappings from view updates into updates on the conceptual schema. Concepts such as virtual updates and three types of insertability are introduced. We also present two algorithms, the View Updatability Algorithm and the View Update Translation Algorithm.

Traditionally, database design activities are partitioned into distinct phases in which a logical design phase precedes physical database design. The objective of the logical design step is to eliminate redundancies and updating anomalies using the notion of data dependencies, while leaving the physical design step to consider how the database schema may be restructured to provide more efficient access. We argue in this paper that the separation of these two steps often results in physical database design not being able to benefit from knowledge of the semantics of data captured in the earlier phases of the database design life cycle. As a step towards overcoming this problem, we demonstrate how classical functional dependencies can be extended to capture data semantics relevant to the design of database schemas which are more desirable from the efficiency point of view. This is accomplished via the introduction of strong and weak functional dependencies. Strong functional dependencies indicate that the relationship between two attributes almost never change. This concept allows us to have controlled redundancies which is beneficial as it can reduce dramatically the effort needed to access frequently needed information. Weak functional dependencies capture the common situations in real life where classical functional dependencies between two attributes hold in general but may be violated in rare cases. Three new normal forms — the relaxed 3NF, replicated 3NF and relax-replicated 3NF, induced by the strong and weak functional dependencies, provide a theoretical framework for designing database schemas which are more efficient and practical, while not compromising the integrity of the underlying database. That is, relations in these new normal forms will not suffer from undesirable updating anomalies.

The increasingly large amount of personal information poses a critical problem to users. Traditional file organization in hierarchical directories is not suited to the effective management of personal information. In order to overcome the shortcomings of the current hierarchical file system and efficiently organize and maintain personal information, some new tools are expected to be invented. In this paper, we propose a novel scheme called concept space - a network of concepts and their associations – and use topic map as the underlying data model. We present a materialized view scheme to provide users with a flexible view of the file system according to their own cognition. We also reduce the storage requirement to save space usage of this system by borrowing some ideas from XML data management and contriving a novel and efficient data compression scheme. To demonstrate the effectiveness of the above idea, we have implemented a prototype personal information management system called NovaPIM and presented its system architecture. Extensive experiments show that our proposed scheme is both efficient and effective.

We develop a logic for entity-relationship databases, ERL, that is a generalization of database logic. ERL provides advantages to the ER model much as FOL (first-order logic) does to the relational model: a uniform language for expressing database schema, integrity constraints, and database manipulation; clearly defined semantics; the capability to express database transformations; and deductive capabilities. We propose three query languages for ER databases called ERRC, ERSQL, and ERQBE, which are generalizations of the relational calculus, SQL, and QBE, respectively. We use example queries and updates to demonstrate the capabilities of these languages. We apply database transformations to introduce the notion of views and to show that both ERRC and ERSQL are relationally complete.

Users often do not require a complete answer to their query but rather only a sample. They expect the sample to be either the largest possible or the most representative (or both) given the resources available. We call the query processing techniques that deliver such results 'approximate'. Processing of queries to streams of data is said to be 'progressive' when it can continuously produce results as data arrives. In this paper, we are interested in the progressive and approximate processing of queries to data streams when processing is limited to main memory. In particular, we study one of the main building blocks of such processing: the progressive approximate join. We devise and present several novel progressive approximate join algorithms. We empirically evaluate the performance of our algorithms and compare them with algorithms based on existing techniques. In particular we study the trade-off between maximization of throughput and maximization of representativeness of the sample.