Networks, such as social networks, biochemical and protein-protein interaction networks are ubiquitous in the real world. Network representation learning aims to embed nodes a network low-dimensional, dense, real-valued vectors, facilitate downstream analysis. The existing embedding methods commonly endeavor capture structure information network, but lack of consideration subsequent tasks synergies between these tasks, which equal importance for desirable representations. To address this...

Abstract Data mining is a process to extract unknown, hidden, and potentially useful information from data. But the problem of data island makes it arduous for people collect analyze scattered data, there also privacy security issue when A collaboratively decentralized approach called federated learning unites multiple participants generate shareable global optimal model keeps privacy‐sensitive on local devices, which may bring great hope us solving problems protection. Though has been...

Abstract Breast cancer is the second deadliest among women. Mammography an important method for physicians to diagnose breast cancer. The main purpose of this study use deep learning automatically classify masses in mammograms into benign and malignant. This proposes a two‐view classification model consisting convolutional neural network (CNN) recurrent (RNN), which used malignant masses. composed two branch networks, modified ResNet are extract breast‐mass features from craniocaudal (CC)...

Objectives: Our study aims to investigate the associations between sexual minority stressors, resilience factors, and substance misuse outcomes, using an intersectional framework examine heterogeneities across populations (SMPs). The hypothesized that factors would mitigate adverse effects of stressors on alcohol drug risks. Design: current employed a secondary data analysis strategy analyze cross sectional Bayesian hierarchical modeling multilevel individual heterogeneity discriminatory...

Objectives/Goals: Sexual minority populations (SMPs), including lesbian, gay, and bisexual groups, disproportionately encounter discriminatory experiences due to bi/homonegativity systemic inequities across various social domains. We aim understand how the neighborhood-level stressors resilience sources differed specific groups in SMPs. Methods/Study Population: Utilizing NIH All of Us’ cloud-based platform, we selected cohorts self-identifying as gay (n = 9,454), 15,284), lesbian 5267), or...

Objectives/Goals: Discriminatory experiences within healthcare settings significantly hinder equitable health access for sexual minority groups (SMPs) in the USA. These discriminatory can manifest various forms (e.g., refusal of care). We aimed to explore different types discrimination encountered by SMPs settings. Methods/Study Population: This study utilized secondary data from NIH All Us Research Program. For this analysis, we selected cohorts self-identifying as gay (n = 9,454), bisexual...

Objectives/Goals: Sexual minority populations report a disproportionately high prevalence of alcohol use, often attributed to coping with bi/homonegativity and systemic inequities across various social domains. This study aims explore use patterns associated neighborhood individual factors among sexual (SMPs) using data from the NIH All US dataset. Methods/Study Population: Alcohol was assessed AUDIT-C (Alcohol Use Disorders Identification Test—Consumption) scale sample 9,454 gay, 15,284...

Recently, multi-task learning (MTL) has been extensively studied for various face processing tasks including detection, landmarks localization, pose estimation and gender recognition, which endeavors to train a better model by exploiting the synergy among related tasks. However, raw dataset used training often contains sensitive private information, can be maliciously recovered carefully analyzing outputs. To address this problem, we propose novel privacy-preserving approach, utilizes...

Real-world networks are composed of diverse interacting and evolving entities, while most existing researches simply characterize them as particular static networks, without consideration the evolution trend in dynamic networks. Recently, significant progresses tracking properties have been made, which exploit changes entities links network to devise embedding techniques. Compared widely proposed methods, endeavors encode nodes low-dimensional dense representations that effectively preserve...

The semantic Web has enabled the creation of a growing number knowledge bases (KBs), which are designed independently using different techniques. Integration KBs attracted much attention as usually contain overlapping and complementary information. Automatic techniques for KB integration have been improved but far from perfect. Therefore, in this paper, we study problem base crowd intelligence. There both classes instances KB, our work, propose novel hybrid framework considering...

Multi-task learning (MTL) is a paradigm which can improve generalization performance by transferring knowledge among multiple tasks. Traditional collaborative filtering recommendation methods suffer from cold start, sparsity and scalability problems. The latest research has shown that applying side information of graph not only solve the problems above, but also accuracy recommendation. However, existing multi-task for enhanced expose obvious issues disclosing private training samples. In...

Objectives The application of artificial intelligence (AI) to the field pathology has facilitated development digital pathology, hence, making AI-assisted diagnosis possible. Due variety lung cancers and subjectivity manual evaluation, invasive non-mucinous adenocarcinoma (ADC) is difficult diagnose. We aim offer a deep learning solution that automatically classifies ADC histological subtypes. Design For this investigation, 523 whole-slide images (WSIs) were obtained. divided 376 WSIs at...

The Internet of Things (IoT) aims to achieve the interconnection all devices in our lives. Due complex network environment, IoT with mobile often faces many security problems, such as privacy leakages and identity forgery attacks. As a developing technology IoT, near field communication (NFC) is widely used electronic payments authentications. current NFC studies mainly focus on payment technology, but there are few protection lightweight requirements authentication protocol. We IoT. In...

To improve the test automation in software development process, following researches on cases generation technology from models, an incremental case approach is proposed based finite automata, and Event deterministic automata (ETDFA) are employed to describe sequence diagram models of system interaction. By model checked with Propositional projection temporal logic (PPTL), correctness ETDFA verified. Then we can get composed by synthesis rules, generate incrementally algorithm. Case studies...

With the development of Internet, increase information sources and speed release transmission have led to a sharp in amount information. To enable users finding more accurate reliable large heterogeneous multi-source data, data fusion technology becomes important. Data structuralizes integrates from different which greatly improves comprehensiveness, availability extensibility data. This paper proposes general framework. The framework transforms structured semi-structured unstructured into...

Abstract The Poisson distribution arises naturally when dealing with data involving counts, and it has found many applications in inverse problems imaging. In this work, we develop an approximate Bayesian inference technique based on expectation propagation for approximating the posterior formed from likelihood function a Laplace type prior distribution, e.g. anisotropic total variation prior. approach iteratively yields Gaussian approximation, at each iteration, updates approximation to one...