Search

BACKGROUND: CLOVES (congenital lipomatous overgrowth, vascular malformations, epidermal nevi, scoliosis/skeletal/spinal) syndrome is a rare and progressive genetic disorder resulting from somatic mosaicism in activating mutations in the phosphatidylinositol-4,5- bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) gene. PIK3CA is a cell growth master regulator where gain of function mutations give rise to abnormal activation of the PI3K-AKT- mammalian target of rapamycin (mTOR) pathway. Treatment with sirolimus, an mTOR inhibitor, may therefore be of benefit in patients with CLOVES syndrome. OBJECTIVE: Here we describe the efficacy and toxicity of sirolimus in a pediatric patient with progressive CLOVES syndrome. RESULTS: The child presented with a large and painful abdominal malformation, massive overgrowth of his feet, limb length discrepancy and genu valgum. There was dramatic clinical and radiographic improvement in the size and comfort of his abdominal mass within several months of initiating medical therapy. This, combined with orthopaedic care of his genu valgum, leg length discrepancy, and overgrowth of his feet, has allowed for significant functional gains. CONCLUSIONS: Multidisciplinary care is essential for comfort and functional gains in patients with CLOVES syndrome, particularly those with severe symptoms. Close monitoring while on sirolimus medical therapy combined with frequent reassessment of orthopedic needs can dramatically improve patient quality of life and outcomes.

The purpose of this systematic review is to evaluate for evidence of the association of acupuncture with relieving the symptoms of adenomyosis. We searched ten electronic databases and included randomized controlled trials (RCTs) in women with adenomyosis. The methodological quality was moderate evidence level by Cochrane risk-of-bias criteria. The results were analyzed by Review Manager 5.3 and expressed as standardized mean differences (SMD) or mean differences (MD). Eleven RCTs (942 subjects) were included in this meta-analysis. Analysis with no heterogeneity showed that acupuncture group obtained a significant better effect (95% CI, -0.48 to -0.10; I2 = 0%) on reducing the size of the uterus and was superior in the shrink of carbohydrate antigen 125 (CA125) level (95% CI, -1.13 to -0.44; I2 = 0%) than that in pharmacological medicine alone group. Moreover, acupuncturewas significantly associated with improving patients' dysmenorrheal symptoms after 3-month menstrual cycles treatment (95% CI, -1.25 to -0.13). The adverse events, especially the incidence of hot flashes, -1.25 to -0.13). The adverse events, especially the incidence of hot flashes, were less reported in acupuncture group compared to pharmacological medicine alone group (Odds Ratio, OR, 0.17; 95% CI, 0.08 to 0.35; I2 = 0%). Acupuncture therapy is a promising avenue for the development of alternatives to surgery and medicine in the treatment of adenomyosis. However, further rigorous trials are needed to confirm the claims of our results.

Currently, the Newton-Raphson iterative algorithm has been extensively employed in the fields of basic research and engineering. However, when noise components exist in a system, its performance is largely affected. To remedy shortcomings that the conventional computing methods have encountered in a noisy workspace, a novel modified Newton integration (MNI) algorithm is proposed in this article. In addition, the steady-state error of the proposed MNI algorithm is smaller than that of the Newton-Raphson algorithm under a noise-free or noisy workspace. To lay the foundations for the corresponding theoretical analyses, the proposed MNI algorithm is first converted into a homogeneous linear equation with a residual term. Then, the related theoretical analyses are carried out, which indicate that the MNI algorithm possesses noise-tolerance ability under various noisy environments. Finally, multiple computer simulations and physical experiments on robot control applications are performed to verify the feasibility and advantage of the proposed MNI algorithm.

Warhead power assessment of the anti-ship missile plays a vital role in determining the optimal design of missile, thus having important strategic research significance. However, in the assessment process, expert’s judgement will directly affect the assessment accuracy. In addition, there are many criteria involved in the missile design alternatives. Some criteria with poor performance may be compensated by other criteria with excellent performance, and then it is impossible to find the truly optimal alternative. Aimed at solving these problems, this paper proposes a synthetical assessment process based on fuzzy hesitant linguistic term set and the Gained and Lost Dominance Score (GLDS) method. In order to improve the assessment accuracy of experts and solve the problem that experts generate different opinions, combined with the advantages of fuzzy hesitant sets and linguistic term sets, the double hierarchy hesitant fuzzy linguistic term sets are used in this paper to improve the accuracy of expert’s judgement. In order to effectively combine expert’s experience with the data of criteria, the evidence theory and entropy weight method are used to transfer the expert’s judgement to the weight. In order to avoid selecting defective alternative of missile design, the GLDS is used to fuse expert information and criteria information. Sensitivity analysis shows that the assessment process has sensitivity to some extent. However, when the fluctuation of expert’s assessment makes the fluctuation of θ in the range of −5% to 5%, the impact on the results is not quite conspicuous. The analysis of calculation result and comparative analysis show that the assessment process proposed in this paper is accurate enough, has great advantage in selecting the current and potential optimal alternative of missile design, and avoids the alternatives with low criteria performance that cannot be compensated by other criteria being selected.

Orthognathic surgery is a procedure that consists of correcting the abnormalities of the maxilla, mandible or both. Maxillary movements have a direct effect on the upper airways. For this reason, in the preoperative period, bone movements should be carefully analyzed and planned to avoid causing any damages to the airways. The goal of this study was to discuss the long-term evolution of the airways after orthognathic surgery. In order to do that, a search was performed through the databases PubMed, Virtual Health Library (VHL) and Scopus, using the acronym PICO strategy and a guiding question to conduct the search. Through the keywords and boolean operators "and" and "or" we found 930 articles. After reading the titles, abstracts and application of inclusion and exclusion criteria, we had 21 articles that were studied in their integrity. At the end of the selection, only 8 studies were included in the review. The postoperative evaluations from the selected studies showed that the airways suffered variations, independently of the patient's facial pattern. In addition, the airways tend to return to initial volumes, but none of the included studies obtained a complete return to the initial airways, even 6 years after surgery. Therefore, more studies with controlled methodologies should be conducted in order to have a better understanding of the airways and their changes.

Multi-task optimization aims to solve two or more optimization tasks simultaneously by leveraging inter-task knowledge transfer. However, as the number of tasks increases to the extent of many-task optimization, the knowledge transfer between tasks encounters more uncertainty and challenges, thereby resulting in degradation of optimization performance. To give full play to the many-task optimization framework and minimize the potential negative transfer, this paper proposes an evolutionary many-task optimization algorithm based on a multi-source knowledge transfer mechanism namely EMaTO-MKT. Particularly, in each iteration, EMaTO-MKT determines the probability of using knowledge transfer adaptively according to the evolution experience, and balances the self-evolution within each task and the knowledge transfer among tasks. To perform knowledge transfer, EMaTO-MKT selects multiple highly similar tasks in term of maximum mean discrepancy as the learning sources for each task. Afterward, a knowledge transfer strategy based on local distribution estimation is applied to enable the learning from multiple sources. Compared with the other state-of-the-art evolutionary many-task algorithms on benchmark test suites, EMaTO-MKT shows competitiveness in solving many-task optimization problems.

This paper focuses on communication-efficient federated learning problem, and develops a novel distributed quantized gradient approach, which is characterized by adaptive communications of the quantized gradients. Specifically, the federated learning builds upon the server-worker infrastructure, where the workers calculate local gradients and upload them to the server; then the server obtain the global gradient by aggregating all the local gradients and utilizes it to update the model parameter. The key idea to save communications from the worker to the server is to quantize gradients as well as skip less informative quantized gradient communications by reusing previous gradients. Quantizing and skipping result in 'lazy' worker-server communications, which justifies the term Lazily Aggregated Quantized (LAQ) gradient. Theoretically, the LAQ algorithm achieves the same linear convergence as the gradient descent in the strongly convex case, while effecting major savings in the communication in terms of transmitted bits and communication rounds. Empirically, extensive experiments using realistic data corroborate a significant communication reduction compared with state-of-the-art gradient- and stochastic gradient-based algorithms.

Data-driven models provide upgradeable environments to simulate inflow to reservoirs. An important aspect in the development of an adaptive neural fuzzy inference system (ANFIS) model is to choose the correct procedure (i.e. efficient and effective) to train the model. In this study, a daily timescale ANFIS-based model was developed to simulate aggregated monthly long-term inflow to the Ross River reservoir in northern Queensland, Australia. The suitability of different evolutionary algorithms (EAs) was evaluated to train an ANFIS-based model, including a genetic algorithm (GA), particle swarm optimisation (PSO), shuffled frog leaping algorithm, biogeography-based optimisation, harmony search algorithm, differential evolution algorithm, invasive weed optimisation (IWO) and a cultural algorithm. Four indices – the root mean square error, Nash–Sutcliffe efficiency, reliability index and vulnerability index – were used to measure the performance of the model. It was found that application of either IWO, GA or PSO provided accurate simulated inflow time series. The outputs obtained using the other EAs were not sufficiently accurate. Use of a coupled EA–ANFIS-based model was found to improve the accuracy of simulating long-term monthly inflow compared with other models used in a few previous recent studies.

The purpose of this study was to compare the learning curves and radiographic variables between screw and suture button fixation of the allograft for surgeons learning the all-arthroscopic anatomic glenoid reconstruction (AAGR) technique for anterior shoulder instability. We compared the surgical times of these two fixation techniques, as well as the graft placement in the vertical and medial-lateral directions.This was a retrospective review of patients who underwent AAGR for recurrent anterior shoulder instability. Start and end times were recorded for each procedure, and learning was assessed through the change in operative time over successive surgeries and by variability of operative time. Graft placement, in terms of vertical positioning, medial-lateral step formation, and obliquity of fixation (alpha angle), was evaluated using 3-dimensional CT scans at approximately 6 months postoperation.A total of 43 patients were included in our study. Twenty-seven had screw fixation, and 16 had suture button fixation. The surgical duration of the button fixation technique was comparable to that of screw fixation (Suture button fixation of the allograft is a potential alternative to screw fixation for the management of glenoid bone loss in recurrent anterior shoulder instability. The results of this study show that this method has a comparable learning curve, with a similar surgical duration, while not compromising the optimal accuracy of vertical and medial-lateral placement of the graft and achieves a superior alpha angle of fixation.Level III, retrospective cohort study.

Background: The worldwide prevalence of eczema has continued to rise over the past decades.This has led to emphasis on the association between air pollution and eczema. This study investigated the relationship between daily exposure to air pollution and the number of eczema outpatient visits in Guangzhou with the overarching goal of providing novel insights on the interventions for eczema aggravation and prevention.Methods: Daily air pollution data, meteorological data, and number of eczema outpatients were obtained from 18th January 2013 to 31st December 2018 in Guangzhou. A generalized additive model with Poisson distribution was used to assess the association between the number of eczema outpatient visits and short-term exposure of PM2.5 and PM10. In addition, the effect of PM2.5 and PM10 by age (Results: A total of 293,343 eczema outpatient visits were recorded. The obtained results indicated that a 10μg/m3 increase of the same day/ lag 1 day/ lag 2 days PM2.5 was associated with increments of 2.33%(RR=1.0233, 95%CI: 1.0206-1.0260, PPP3 increase of PM10 was associated with eczema outpatients risk increments of 1.97%(RR=1.0197, 95%CI: 1.0177-1.0217, PPP2.5 exposure and eczema were observed at lag 0 day with the percent changes being 4.72% (95%CI, 4.18-5.28%, PPPConclusion: Short-term exposure of PM2.5 and PM10 increases the number of eczema outpatients especially among children and the elderly. This study has provided a further understanding of the relationship between air pollutants and eczema, which will benefit disease prevention and lower the health burden.

In the field of Deep Learning, the long short-term memory (LSTM) networks lie in the category of recurrent neural network (RNN) architectures. The distinctive capability of the LSTM is learning non linear long term dependency structures. This makes the LSTM a good candidate for prediction tasks in non linear time dependent systems such as prediction of runoff in a catchment. In this study, we use a large sample of 740 gauged catchments with very diverse hydro-geo-climatic conditions across France. We present a regime classification based on three hydro-climatic indices to identify and classify catchments with similar hydrological behaviors. We do this because we aim to investigate how regime derived information can be used in training LSTM-based runoff models. The LSTM-based models that we investigate include local models trained on individual catchments as well as regional models trained on a group of catchments. In local training, for each regime, we identify the optimal lookback, i.e. the length of the sequence of past forcing data that the LSTM needs to work through. We then use this length in training regional models that differ in two aspects: 1) hydrological homogeneity of the catchments used in their training, 2) configuration of the static attributes used in their inputs. We examine how each of these aspects contributes to learning of the LSTM in regional training. At every step of this study, we benchmark performances of the LSTM against a conceptual model (GR4J) on both train and unseen data. We show that the optimal lookback is regime dependent and homogeneity of the train catchments in regional training has a more significant contribution to learning of the LSTM than the number of the train catchments.

ABSTRACT INTRODUCTION The prevalence of long-term health conditions (LTCs) continues to increase and it is normal for people to have several. Lifestyle is a core feature of the self-management support given to people with LTCs, yet it seems to fail to meet their needs. From a larger study exploring the experiences of this group, this paper reports on the role of food and mealtimes, and the effect of the nutritional advice on the lives of people with several LTCs. METHODS The experiences of 16 people with several advanced LTCs were explored using multiple qualitative case studies. Over an 18-month period, contact with participants included monthly conversations, two longer, semi-structured interviews and one interview with their primary care clinicians. The data were analysed inductively using thematic analysis. RESULTS There was profound social, psychological and cultural significance of food to the participants. These aspects of food were then described using the Whare Tapa Wha model of health. The approach to food and nutrition participants experienced in health care was closely associated with conventional weight management. They found this advice largely irrelevant and therefore ineffective. DISCUSSION The pressure people with several LTCs experience during clinical encounters has perverse effects on wellbeing and reinforces an overall sense of failure. This research challenges the assumption that all people have the will, agency and income to follow the advice they receive about nutrition and exercise. Considering people’s needs using the Whare Tapa Wha model of health offers clinicians a framework to re-conceptualise the core features of self-management support and provide realistic, compassionate care that will optimise the quality of life for people with multiple LTCs.

Wafer map defect recognition (WMDR) is an important part of the integrated circuit manufacturing system. Although deep neural networks (DNNs) have been applied successfully in WMDR, class imbalance and lack of data with class labels affect their performance significantly. A new generative adversarial network (GAN), multi-granularity generative adversarial network (MGGAN), is proposed for wafer map augmentation and enhancement. To alleviate instability and mode collapse of traditional GANs, the lightweight convolution and a two-way information interaction of three subnetworks are considered. MGGAN consists of an auxiliary feature extractor (AFE), a generator (G) and a discriminator (D) for wafer map generation and WMDR. Firstly, a pretrained deep convolutional neural network (CNN), ResNet101, is employed as AFE to extract multi-granularity features from wafer maps. Secondly, in order to improve effectiveness of the adversarial training, a feature matching term is considered in the objective function of the feature generator to minimize the statistical difference between the real images and the generated images. Finally, MGGAN has been successfully applied to WMDR. The experiment results on an industrial dataset WM-811K demonstrate that MGGAN outperforms other typical GANs that aim to solve class imbalance problems and gains better recognition performance than those state-of-the-art DNNs on WMDR. This indicates effectiveness of MGGAN in image enhancement and generation. MGGAN obtains an accuracy of 88.02% on the original data and the pre-trained ResNet101 obtains 93.43% on the data enhanced by MGGAN.

Text recognition in natural scene images endures as a challenging problem, attributable to a highly variable appearance in an unconstrained environment. The proposed work is novel of its kind, which improved the generalization of the Twin Support Vector Machine (T-SVM) by manifold regularization, that was further extended to validate and recognize the text in the natural scene images. The multi-class T-SVM is augmented with ambient regularizer term and intrinsic regularizer term, assisting to form a smooth function for the model. Text comprehension from natural scene includes the localization, recognition, and reconstruction of the text. The work includes an additional module, revalidation that discards the false positives from the text objects detected during the localization phase. Then in recognition phase, each letter from the pool of text objects is identified for the appropriate class and provided as output to the text construction phase, which builds the text using the pool of coordinates associated with the object. The model is evaluated against traditional methods like Support Vector Machine (SVM), T-SVM, LST-SVM(Least Square Twin Support Vector Machine) and also with other research of the same kind and shows the accuracy of 84.91% with ICDAR 2015, 84.21% with MSRA 500 and 86.21% with SVT. The analyses of the results show that the model was capable of recognizing most of the characters from the image along with consummate accuracy level.

Deep neural networks are vulnerable to adversarial attacks. More importantly, some adversarial examples crafted against an ensemble of source models transfer to other target models and, thus, pose a security threat to black-box applications (when attackers have no access to the target models). Current transfer-based ensemble attacks, however, only consider a limited number of source models to craft an adversarial example and, thus, obtain poor transferability. Besides, recent query-based black-box attacks, which require numerous queries to the target model, not only come under suspicion by the target model but also cause expensive query cost. In this article, we propose a novel transfer-based black-box attack, dubbed serial-minigroup-ensemble-attack (SMGEA). Concretely, SMGEA first divides a large number of pretrained white-box source models into several ``minigroups.'' For each minigroup, we design three new ensemble strategies to improve the intragroup transferability. Moreover, we propose a new algorithm that recursively accumulates the ``long-term'' gradient memories of the previous minigroup to the subsequent minigroup. This way, the learned adversarial information can be preserved, and the intergroup transferability can be improved. Experiments indicate that SMGEA not only achieves state-of-the-art black-box attack ability over several data sets but also deceives two online black-box saliency prediction systems in real world, i.e., DeepGaze-II (https://deepgaze.bethgelab.org/) and SALICON (http://salicon.net/demo/). Finally, we contribute a new code repository to promote research on adversarial attack and defense over ubiquitous pixel-to-pixel computer vision tasks. We share our code together with the pretrained substitute model zoo at https://github.com/CZHQuality/AAA-Pix2pix.

Background: Macrodactyly is a rare congenital disorder of overgrowth affecting the digits of the upper or lower extremity. Mostly, patients are surgically treated during childhood to reduce the digit or to stop growth. There are no standardized guidelines for the treatment and follow-up of macrodactyly. Consequently, follow-up may not be regularly scheduled into adulthood. Methods: A retrospective, descriptive analysis of patients with the long-term progression of macrodactyly who presented at our tertiary referral hospital between July 2018 and March 2020 was performed. All patients from our local macrodactyly database were screened for progression of macrodactyly since adulthood; this resulted in four patients. The aim of these case series is to highlight the clinical features and disease course at long-term follow-up. Results: All patients were surgically treated during childhood and showed progression of tissue overgrowth during adult life. All patients developed severe secondary degenerative bone changes in macrodactyly affected digits, such as ankyloses of joints, new bone formation, and bony spurs. Subsequently, tissue overgrowth and degenerative bone changes led to functional problems. Conclusion: Patients with macrodactyly may experience growth during adult life, which may progress to deforming changes. Consequently, patients should be informed about the possible growth, and the progressive growth should be monitored.

Due to the potential hazards of traversing the field ground, the robotic ground classification (RGC) has been widely concerned and extensively studied in environmental perception tasks. However, the RGC usually performs well in experimental environment (source domain) but poorly in working environment (target domain), as the environmental change might cause the ground properties variation and consequently the data drift that comes down to the domain adaptation problem. Hence, we propose the broad feature alignment (BFA) to suppress the deterioration in accuracy of RGC upon dynamic environment. The contribution is threefold: (i) the features are represented first via a broad learning network to improve the feature alignment performance; (ii) the target domain information preserving term is introduced to design a specific alignment object for suppressing data drift, so that the source-domain broad features could be aligned to the designed alignment object via projected maximum mean discrepancy (PMMD); (iii) the feature-temporal manifold regularizer is exploited to improve the alignment consistency of source-domain represented features. The proposed method is verified experimentally on the data gathered by a micro tracked robot.

Segmentation of MRI images in the presence of artifacts such as noise and bias field effect is challenging. In this work, to handle both artifacts, we have formulated an intuitionistic fuzzy set theory based bias-corrected intuitionistic fuzzy c-means with spatial neighborhood information method for MRI image segmentation. We represent the image in terms of intuitionistic fuzzy sets using two well-known intuitionistic fuzzy generation functions, namely Sugeno's and Yager's negation functions. In this way, the proposed approach takes advantage of the intuitionistic fuzzy set theory to address the uncertainty and vagueness present in data. The spatial neighborhood information term helps to retain the fine image details in the segmentation process, which is usually lost when smoothing is used. The bias field effect is modeled as a slow varying intensity field additive in nature with mean zero in IFS framework. The proposed approach estimates the bias field and produces a segmented image simultaneously. The segmentation performance of the proposed method is evaluated in terms of the Dice score, average segmentation accuracy, and index of variation. Further, the comparison of the proposed method with other similar state-of-the-art methods on two well-known publicly available Brain MRI datasets shows the significant improvements in segmentation performance in terms of average segmentation accuracy, Dice score, and index of variation. The proposed approach achieves 93% average segmentation accuracy in the presence of severe bias field and noise artifacts on BrainWeb simulated dataset, which outperformed the state-of-the-art methods.

The aim of this study was to determine the short-term compliance with regular home monitoring of macular retinal sensitivity (RS) in intermediate age-related macular degeneration (iAMD). Home-based outcomes were compared with in-clinic outcomes determined using (1) the same tablet device under supervision, and (2) the Macular Integrity Assessment (MaIA) microperimeter.Single-center longitudinal compliance and reliability study.A total of 73 participants with iAMD were trained to perform macular field testing with the Melbourne Rapid Fields-macular (MRF-m) iPad application. Volunteers were asked to return 6 weekly tests from home, guided by audio instructions. We determined compliance with weekly testing and surveyed for factors that limited compliance. Test reliability (false positive, false negative) and RS were compared to in-clinic assays (MaIA). Data are given as mean ± SD or as median [quartile 1-3 range]. Group comparisons were achieved with bootstrap to define the 95% confidence limits.A total of 59 participants submitted 6 home examinations with a median intertest interval of 8.0 [7.0-17] days. Compliance with weekly testing (7 days ±24 hours) was 55%. The main barrier to compliance was information technology (IT) logistic reasons. Of 694 home examinations submitted, 96% were reliable (false-positive results25%). The mean RS returned by the tablet was significantly higher (+3.2 dB, P.05) compared to the MaIA.Home monitoring produces reliable results that differ from in-clinic tests because of test design. This should not affect self-monitoring once an at-home baseline is established, but these differences will affect comparisons with in-clinic outcomes. Reasonable compliance with weekly testing was achieved. Improved IT support might lead to better compliance.

Acute renal failure requiring dialysis after heart transplantation remains a significant clinical issue because of its increasing incidence. We aimed to investigate its time trends, clinical predictors, and long-term outcomes.Adult heart transplantation recipients registered in the United Network for Organ Sharing registry between 2009 and 2020 were identified. The patients were grouped according to the requirement for dialysis in the postoperative heart transplantation period. The independent risk predictors were identified, and the association between post-heart transplantation renal failure requiring dialysis and long-term mortality accounting for re-transplantation was investigated.A total of 28,170 patients were included in the study, of which 3,371 (12%) required dialysis immediately post-heart transplantation. The incidence increased from 7.9% to 13.9% during the study period. Longer ischemic time, serum creatinine at transplantation1.2 mg/dL, prior cardiac surgery, higher recipient body mass index, support of mechanical ventilation or extracorporeal membrane oxygenation, and history of congenital heart disease or restrictive/hypertrophic cardiomyopathy were its predictors (all p0.05). Patients on posttransplant dialysis had a higher risk of all-cause mortality (adjusted hazard ratio [aHR]: 5.2, 95% CI: 4.7-5.7, p0.001), 30 day mortality (aHR: 7.7, 95% CI: 6.3-9.6, p0.001) and 1 year mortality (aHR: 7.5, 95% CI: 6.6-8.6, p0.001). Post-transplant dialysis was associated with a risk of treated rejection at 1 year.Acute renal failure requiring dialysis after heart transplantation is associated with significantly worse 30 day and long-term mortalities, and thus, early identification of high-risk patients is crucial to prevent severe renal complications.

The core component of most anomaly detectors is a self-supervised model, tasked with modeling patterns included in training samples and detecting unexpected patterns as the anomalies in testing samples. To cope with normal patterns, this model is typically trained with reconstruction constraints. However, the model has the risk of overfitting to training samples and being sensitive to hard normal patterns in the inference phase, which results in irregular responses at normal frames. To address this problem, we formulate anomaly detection as a mutual supervision problem. Due to collaborative training, the complementary information of mutual learning can alleviate the aforementioned problem. Based on this motivation, a SIamese generative network (SIGnet), including two subnetworks with the same architecture, is proposed to simultaneously model the patterns of the forward and backward frames. During training, in addition to traditional constraints on improving the reconstruction performance, a bidirectional consistency loss based on the forward and backward views is designed as the regularization term to improve the generalization ability of the model. Moreover, we introduce a consistency-based evaluation criterion to achieve stable scores at the normal frames, which will benefit detecting anomalies with fluctuant scores in the inference phase. The results on several challenging benchmark data sets demonstrate the effectiveness of our proposed method.

Objectives. The aim of this study was to compare the long-term oncologic outcomes of sentinel lymph node biopsy (SLNB) versus elective neck dissection (END) in clinically node-negative (cN0) tongue cancer.Methods. This was a retrospective cohort study of patients with cN0 tongue cancer from a single institution, including 91 patients in the SLNB group and 120 patients in the END group.Results. The overall recurrence rate showed no significant difference between the two groups. The regional control rate was also comparable between the two groups (P=0.490). The 5-year recurrence-free survival (RFS) was slightly better in the SLNB group than in the END group (P=0.427). The 5-year overall survival (OS) rate was 89.9% in the SLNB group versus 91.9% in the END group (P=0.737). In a propensity-matched subgroup analysis, the type of neck management did not affect RFS or OS.Conclusion. SLNB showed non-inferior oncologic outcomes compared to END in patients with cN0 tongue squamous cell carcinoma.

Background: Epidemiological data on anticoagulation for venous thromboembolism (VTE) in prostate cancer are sparse. We aimed to investigate associations between anticoagulation duration and risks of VTE recurrence after treatment cessation and major on-treatment bleeding in men with prostate cancer in Sweden. Methods: Using nationwide prostate cancer registry and prescribing data, we followed 1413 men with VTE and an outpatient anticoagulant prescription following prostate cancer diagnosis. Men were followed to identify cases of recurrent VTE, and hospitalized major bleeding. We calculated adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) to quantify the association between anticoagulation duration (reference ≤3 months) and recurrent VTE using Cox regression. We estimated 1-year cumulative incidences of major bleedings from anticoagulation initiation.Results: The outpatient anticoagulation prescribed was parenteral (64%), direct oral anticoagulant (31%), and vitamin K antagonist (20%). Median duration of anticoagulation was 7 months. Adjusted HRs (95% CI) for off-treatment recurrent pulmonary embolism (PE) were 0.32 (0.09–1.15) for >3–6 months’ duration, 0.21 (0.06–0.69) for >6–9 months and 0.16 (0.05–0.55) for >9 months; corresponding HRs for deep vein thrombosis (DVT) were 0.67 (0.27–1.66), 0.80 (0.31–2.07), and 1.19 (0.47–3.02). One-year cumulative incidences of intracranial, gastrointestinal and urogenital bleeding were 0.9%, 1.7%, 3.0% during treatment, and 1.2%, 0.9%, 1.6% after treatment cessation.Conclusion: The greatest benefit in reducing recurrent VTE risk was with >9 months anticoagulation for PE and >3–6 months for DVT. Risks of major bleeding were low not being substantially different to the risks after treatment cessation.

One of the key tasks for an intelligent visual surveillance system is to automatically re-identify objects of interest, e.g., persons or vehicles, from nonoverlapping camera views. This demand incurs the vast investigation of person re-identification (re-ID) and vehicle re-ID techniques, especially those deep learning-based ones. While most recent algorithms focus on designing new convolutional neural networks, less attention is paid to the loss functions, which are of vital roles as well. Triplet loss and softmax loss are the two losses that are extensively used, both of which, however, have limitations. Triplet loss optimizes the model to produce features with which samples from the same class have higher similarity than those from different classes. The problem of triplet loss is that the number of triplets to be constructed grows cubically with training samples, which causes scalability issue, unstable performance, and slow convergence. Softmax loss has favorable scalable property and is widely used for large-scale classification problems. However, since Softmax loss only aims to separate well training classes, its performance for re-ID tasks is not desirable because the model is tested to measure the similarity of samples from unseen classes. We propose the support neighbor (SN) loss, which avoids the limitations of the abovementioned two losses. Unlike triplet loss that is calculated based on triplets, SN loss is derived from K -nearest neighbors (SNs) of anchor samples. The SNs of an anchor are unique, containing more valuable contextual information and neighborhood structure of the anchor, and thus contribute to more stable performance and reliable embedding from image space to feature space. Based on the SNs, a softmax-like separation term and a squeeze term are proposed, which encourage interclass separation and intraclass compactness, respectively. Experiments show that SN loss surpasses triplet and softmax losses with the same backbone network and reaches the state-of-the-art performance for both person and vehicle re-ID using a ResNet50 backbone when combined with training tricks.

Utility companies have widely adopted the concept of health index to describe asset health statuses and choose proper asset management actions. The existing application and research works have been focused on determining the current asset health index based on the current condition data. For long-term preventative asset management, it is highly desirable to predict asset health indices in the next few years, especially for asset classes in which the assets share similar electrical, mechanical and condition degradation characteristics. This important problem has not been sufficiently addressed. This paper proposes a sequence learning based method to predict long-term health indices for power asset classes. A comprehensive data-driven method based on sequence learning is presented and solid tests are conducted based on real utility data. The proposed method revealed superior performance with comparison to other prediction methods.

The potential benefits of autonomous systems are obvious. However, there are still major issues to be dealt with before developing such systems becomes a commonplace engineering practice, with accepted and trustworthy deliverables. We argue that a solid, evolving, publicly available, community-controlled foundation for developing next-generation autonomous systems is a must, and term the desired foundation Autonomics. We focus on three main challenges: (i) how to specify autonomous system behavior in the face of unpredictability; (ii) how to carry out faithful analysis of system behavior with respect to rich environments that include humans, physical artifacts, and other systems; and (iii) how to build such systems by combining executable modeling techniques from software engineering with artificial intelligence and machine learning.

Background overview and rationale We co-developed a multi-component virtual care solution (TtLIVE) for the home mechanical ventilation (HMV) population using the aTouchAway™ platform (Aetonix). The TtLIVE intervention includes (1) virtual home visits; (2) customizable care plans; (3) clinical workflows that incorporate reminders, completion of symptom profiles, and tele-monitoring; and (4) digitally secure communication via messaging, audio, and video calls; (5) Resource library including print and audiovisual material. Objectives and brief methods Our primary objective is to evaluate the TtLIVE intervention compared to a usual care control group using an eight-center, pragmatic, parallel-group single-blind (outcome assessors) randomized controlled trial. Eligible patients are children and adults newly transitioning to HMV in Ontario, Canada. Our target sample size is 440 participants (220 each arm). Our co-primary outcomes are a number of emergency department (ED) visits in the 12 months after randomization and change in family caregiver (FC) reported Pearlin Mastery Scale score from baseline to 12 months. Secondary outcomes also measured in the 12 months post randomization include healthcare utilization measured using a hybrid Ambulatory Home Care Record (AHCR-hybrid), FC burden using the Zarit Burden Interview, and health-related quality of life using the EQ-5D. In addition, we will conduct a cost-utility analysis over a 1-year time horizon and measure process outcomes including healthcare provider time using the Care Coordination Measurement Tool. We will use qualitative interviews in a subset of study participants to understand acceptability, barriers, and facilitators to the TtLIVE intervention. We will administer the Family Experiences with Care Coordination (FECC) to interview participants. We will use Poisson regression for a number of ED visits at 12 months. We will use linear regression for the Pearlin Mastery scale score at 12 months. We will adjust for the baseline score to estimate the effect of the intervention on the primary outcomes. Analysis of secondary outcomes will employ regression, causal, and linear mixed modeling. Primary analysis will follow intention-to-treat principles. We have Research Ethics Board approval from SickKids, Children’s Hospital Eastern Ontario, McMaster Children’s Hospital, Children’s Hospital-London Health Sciences, Sunnybrook Hospital, London Health Sciences, West Park Healthcare Centre, and Ottawa Hospital. Discussion This pragmatic randomized controlled single-blind trial will determine the effectiveness and cost-effectiveness of the TtLIVE virtual care solution compared to usual care while providing important data on patient and family experience, as well as process measures such as healthcare provider time to deliver the intervention. Trial registration ClinicalTrials.gov NCT04180722. Registered on November 27, 2019.

Children born very preterm ( 32 weeks' gestation) are at greater risk of motor impairment and executive/attentional dysfunctions than term-born children; however, little is known about how functional tasks, including walking, may be affected by very preterm birth.How does the gait pattern of preschool-age children born 30 weeks compare with term-born controls under a variety of walking conditions?In this prospective cohort study, children born 30 weeks and at term were assessed at 4.5-5 years' corrected age, blinded to birth group. Four walking conditions were assessed using the GAITRite® system: preferred speed, cognitive dual-task, motor dual-task, and tandem walking. Gait variables analysed included speed, cadence, step length, step time, base of support (BOS), and single and double support time. Spatiotemporal variables were compared between groups using linear regression, adjusting for lower-limb length, corrected age at assessment, and number of trials.224 children (112 30 weeks and 112 term-born) were assessed. Gait variables of children born 30 weeks did not differ from their term-born peers when walking at their preferred speed, except for higher BOS variability (mean difference [MD] = 0.19 cm, 95% confidence interval [CI] 0.10, 0.27, p 0.001). Under the motor dual-task condition, children born 30 weeks walked faster (MD= 3.06 cm/s, 95% CI 0.14, 5.97, p = 0.040), with a longer step length (MD= 1.10 cm, 95%CI 0.19, 2.01, p = 0.018), and a wider BOS (MD= 0.37 cm, 95%CI 0.06, 0.67, p = 0.019). In cognitive dual-task and tandem conditions, children born 30 weeks walked with a wider BOS compared with term-born peers (MD= 0.43 cm, 95%CI 0.05, 0.81, p = 0.028; and MD= 0.30 cm, 95%CI 0.09, 0.51, p = 0.005, respectively).This research highlights the need to consider the walking performance of preschool-age children born 30 weeks under challenging conditions, such as dual-task or tandem walking, when assessing gait patterns and planning interventions.

In this work, we introduce, justify and demonstrate the Corrective Source Term Approach (CoSTA) – a novel approach to Hybrid Analysis and Modeling (HAM). The objective of HAM is to combine physics-based modeling (PBM) and data-driven modeling (DDM) to create generalizable, trustworthy, accurate, computationally efficient and self-evolving models. CoSTA achieves this objective by augmenting the governing equation of a PBM model with a corrective source term generated using a deep neural network. In a series of numerical experiments on one-dimensional heat diffusion, CoSTA is found to outperform comparable DDM and PBM models in terms of accuracy – often reducing predictive errors by several orders of magnitude – while also generalizing better than pure DDM. Due to its flexible but solid theoretical foundation, CoSTA provides a modular framework for leveraging novel developments within both PBM and DDM. Its theoretical foundation also ensures that CoSTA can be used to model any system governed by (deterministic) partial differential equations. Moreover, CoSTA facilitates interpretation of the DNN-generated source term within the context of PBM, which results in improved explainability of the DNN. These factors make CoSTA a potential door-opener for data-driven techniques to enter high-stakes applications previously reserved for pure PBM.

This paper studies the persistent effects of short-term peer exposure on long-run performance in a college setting. I exploit the random assignment of undergraduates to peer groups during a mandatory orientation week and track the students’ performance over four years (until graduation). Assignment to orientation week groups with high levels of peer ability is associated with lower performance during the first year at college and a higher probability of early dropout. These adverse effects are driven entirely by the exposure of low-ability students to high-ability peers. Beyond the first year, exposure to higher peer ability during the orientation week negatively affects selection into the college’s most popular major (business administration) and final grade point average. Taken together, the findings suggest that the composition of short-term peer groups matters for individual choices and long-run performance outcomes. This paper was accepted by Yan Chen, decision analysis.

Timber–concrete composite is an efficient hybrid construction material that exploits the advantageous properties of timber and concrete. The use of shear connectors enables the two dissimilar materials to act together, resulting in an increase in global stiffness as well as load-carrying capacity. As this composite material is becoming increasingly more popular in the construction industry, there is a need to develop analysis tools that have general applicability to timber–concrete composite systems with variations in loading schemes, specimen configurations, materials, and types of shear connectors. One such tool, a generic two-dimensional nonlinear finite element model, is proposed in this paper; it is verified through numerical simulations of six experiment series carried out by other researchers. Good agreement between experimentally observed behaviour and numerical simulations was generally obtained.

Forecasting solar irradiance is of utmost importance in supplying renewable energy efficiently and timely. This paper aims to experiment five variants of recurrent neural networks (RNN), and develop effective and reliable 5-minute short term solar irradiance prediction models. The 5 RNN classes are long-short term memory (LSTM), gated recurrent unit (GRU), Simple RNN, bidirectional LSTM (Bi-LSTM), and bidirectional GRU (Bi-GRU); the first 3 classes are unidirectional and the last two are bidirectional RNN models. The 26 months data under consideration, exhibits extremely volatile weather conditions in Jinju city, South Korea. Therefore, after different experimental processes, 5 hyper-parameters were selected for each model cautiously. In each model, different levels of depth and width were tested; moreover, a 9-fold cross validation was applied to distinguish them against high variability in the seasonal time-series dataset. Generally the deeper architectures of the aforementioned models had significant outcomes; meanwhile, the Bi-LSTM and Bi-GRU provided more accurate predictions as compared to the unidirectional ones. The Bi-GRU model provided the lowest RMSE and highest R2 values of 46.1 and 0.958; additionally, it required 5.25*10-5 seconds per trainable parameter per epoch, the lowest incurred computational cost among the mentioned models. All 5 models performed differently over the four seasons in the 9-fold cross validation test. On average, the bidirectional RNNs and the simple RNN model showed high robustness with less data and high temporal data variability; although, the stronger architectures of the bidirectional models, deems their results more reliable.

Preoperative variables can predict short term left ventricular assist device (LVAD) survival, but predictors of extended survival remain insufficiently characterized.Patients undergoing LVAD implant (2012-2018) in the Intermacs registry were grouped according to time on support: short-term (1 year, n = 7,483), mid-term (MT, 1-3 years, n = 5,976) and long-term (LT, ≥3 years, n = 3,015). Landmarked hazard analyses (adjusted hazard ratio, HR) were performed to identify correlates of survival after 1 and 3 years of support.After surviving 1 year of support, additional LVAD survival was less likely in older (HR 1.15 per decade), Caucasian (HR 1.22) and unmarried (HR 1.16) patients (p0.05). After 3 years of support, only 3 preoperative characteristics (age, race, and history of bypass surgery, p0.05) correlated with extended survival. Postoperative events most negatively influenced achieving LT survival. In those alive at 1 year or 3 years, the occurrence of postoperative renal (creatinine HR MT = 1.09; LT HR = 1.10 per mg/dl) and hepatic dysfunction (AST HR MT = 1.29; LT HR = 1.34 per 100 IU), stroke (MT HR = 1.24; LT HR = 1.42), infection (MT HR = 1.13; LT HR = 1.10), and/or device malfunction (MT HR = 1.22; LT HR = 1.46) reduced extended survival (all p ≤ 0.03).Success with LVAD therapy hinges on achieving long term survival in more recipients. After 1 year, extended survival is heavily constrained by the occurrence of adverse events and postoperative end-organ dysfunction. The growth of destination therapy intent mandates that future LVAD studies be designed with follow up sufficient for capturing outcomes beyond 24 months.

Objective : To provide an overview of the literature on the mid-term outcomes of chimney EVAR (ChEVAR) for the treatment of juxtarenal abdominal aortic aneurysms (JAAA). Methods : Different electronic databases were searched for published articles up to January 2020. The eligibility criteria were studies describing mid- or long-term outcomes of chimney EVAR (mean follow-up at least 1 year) for treatment of JAAA, including more than 10 cases, published in English, and with full text available. The outcomes measure were overall survival rate, target vessel patency, and freedom from reintervention at 3 years. Quality of the included studies was analysed using the MINORS criteria. Pooled effect estimates were analysed using random-effect models and heterogeneity was tested using I2 statistics. Results : Thirteen articles met the inclusion criteria. The included studies described 1,019 patients. According to the quality assessment, methodological quality was moderate to poor. The pooled overall survival, freedom from reintervention, and target vessel patency at 3 year was 81.4 % (95%CI 73.8 to 87.9), 85.7% (95%CI 75.6 to 93.5), and 95.1% (95%CI 89.3 to 98.7) respectively. Conclusion : The results of this review show good to acceptable short and mid-term survival and good mid-term durability, which supports that ChEVAR as a suitable alternative in high-risk JAAA. However, proper patient selection for ChEVAR seems essential to attain good mid-term outcomes, and further large prospective and good quality studies are required to demonstrate its long-term results and enable conclusions on specific determinants for outcome.

Freshwater species are particularly vulnerable to emerging threats linked to climate change because they are often already heavily impacted by habitat destruction, pollution, and exploitation. For many harvested populations of freshwater fish, these combined impacts have been mitigated for decades through stocking with captive-bred individuals. However, stocking may lead to loss of genetic variation, which may be crucial for adaptation under climate change. Exploration of sustainable alternatives is therefore paramount. We used a female-based integral projection model (IPM) to assess the consequences of terminating a long-term stocking programme for a population of landlocked, migratory brown trout Salmo trutta, and to evaluate relative effectiveness of alternative management strategies involving harvest regulations and river habitat improvement. The IPM classified individuals by body size, life history stage, and location relative to a hydropower dam, and was parameterised with 50 yr of individual-based data, supplemented with literature values. Model simulations indicated a strong population decline of 22-29% per year without stocking, much of which was attributed to high harvest mortality. Consequently, drastic reductions in fishing pressure were predicted to be necessary to ensure population viability without stocking. Mitigation measures reducing mortality associated with the hydropower dam or restoring spawning areas could further contribute to population viability when combined with changes in harvest regulations. Our results thus emphasise that large changes in management strategies, such as termination of long-term stocking programmes, require a thorough assessment of potential consequences and alternative mitigation strategies using data and models, or, at the very least, a precautionary approach under consideration of on-going climate change. Migratory salmonid · Salmo trutta · Integral projection model · Harvest · Fishing · Stocking · Dam · Hydropower · Trout

Recently, the term ‘Internet of Things’ (IoT) has garnered great attention. As a trusted, dependable, and decentralized approach, blockchain has already been used in IoT. However, the existing blockchain has a number of drawbacks that prevent it from being used as a generic platform for IoT. The nodes in IoT are heavily resource-limited, especially computing and networking resources. Unfortunately, they are necessary for the blockchain to solve complicated puzzles and propagate blocks. In this paper, we propose agent mining and cloud mining approaches to solve the above problem in the blockchain-enabled IoT. To be specific, miners act as mining agents for nodes in IoT, offload mining tasks to cloud computing servers, and use networking resources dynamically. Furthermore, in order to enhance the performance, the access selection of users, computing resources allocation, and networking resources allocation are formulated as a joint optimization problem. We then propose a dueling deep reinforcement learning approach to address this problem. Numerical results justify the effectiveness of our proposed scheme.

In the current work we have used the deep learning based long short-term memory model to predict the strength and peak time of solar cycle 25 by employing the monthly smoothed sunspot number data obtained from WDC-SILSO, Royal Observatory of Belgium, Brussels. We have used the stacked LSTM forecasting model to predict the upcoming cycle 25. From our analysis it has been shown that our proposed model is capable of capturing long term dependencies as well as trend within the data. For cycle 20 and 21 the error difference between predicted as well as observed peak value is 2.3 and 0.7 respectively while the peak prediction error is 1.47% and 0.30%. The RMSE of the model for cycle 20 and 21 is 3.97 and 4.34 respectively. For cycle 22, the AE and RE is 4.6 and 2.16% while the RMSE of the model for this case is 4.50. The predicted peak amplitude of solar cycle 23 and 24 from our proposed model has a relative error of 1.75% and 1.99% respectively from the observed value while the RMSE is 3.4 for cycle 23 and 4.2 for cycle 24. Our projected prediction of cycle 25 using the proposed LSTM model, says that it will be stronger than cycle 24 and weaker than cycle 23. The solar cycle 25 will peak with an amplitude of sunspot number at 171.9 ± 3.4 and will be 47 % stronger than cycle 24. The solar cycle 25 will reach its peak in August 2023 ± 2 months.

Smart Agriculture (SA) incorporates low cost and low energy consuming sensors and devices to enhance quantitative and qualitative agricultural production. However, these devices uses an open channel i.e., Internet and generates large amount of data in real-time and thus, has the potential to be misused. As a consequence, the major concern in the implementation of SA is minimizing the risk of security and data privacy violation (e.g., adversaries performing inference attacks). To address these challenges, we propose PEFL, a deep privacy-encoding based Federated Learning (FL) framework that adopts a perturbation-based encoding and Long-Short Term Memory-AutoEncoder (LSTM-AE) technique to achieve the target of privacy. Then a FL-based Gated Recurrent Unit Neural Network algorithm (FedGRU) is designed using the encoded data for intrusion detection. The experiment results based on ToN-IoT dataset reveals that the PEFL can efficiently identify normal and attack patterns after transformation over other non-FL and FL methods.

The huge size of deep neural networks makes it difficult to deploy on the embedded platforms with limited computation resources directly. In this paper, we propose a novel trimming approach to determine the redundant parameters of the trained deep neural network in a layer-wise manner to produce a compact neural network. This is achieved by minimizing a nonconvex sparsity-inducing term of the network parameters while maintaining the response close to the original one. We present a proximal iteratively reweighted method to resolve the resulting nonconvex model, which approximates the nonconvex objective by a weighted l1 norm of the network parameters. Moreover, to alleviate the computational burden, we develop a novel termination criterion during the subproblem solution, significantly reducing the total pruning time. Global convergence analysis and a worst-case O(1/k) ergodic convergence rate for our proposed algorithm is established. Numerical experiments demonstrate the proposed approach is efficient and reliable.

In this article, members of the American College of Radiology Committee on Drugs and Contrast Media propose a new term for symptoms reported after intravascular exposure to gadolinium-based contrast agents-Symptoms Associated with Gadolinium Exposure, or SAGE. This term is advocated in lieu of other proposed nomenclature that presumes a causal relationship that has not yet been scientifically verified. The purpose of this new term, SAGE, is to assist researchers and clinical providers in describing such symptoms without prematurely causally attributing them to a disease and to standardize reporting of these symptoms to allow for coherent interpretation of related studies.

Deep learning has achieved good performance in short-term traffic forecasting recently. However, the stochasticity and distribution imbalance are main characteristics to traffic flow, and these will bring the uncertainty and induce the network overfitting problem during deep learning. To deal with the problems, a new end-to-end hybrid deep learning network model, named M-B-LSTM, is proposed for short-term traffic flow forecasting in this paper. In the M-B-LSTM model, an online self-learning network is constructed as a data mapping layer to learn and equalize the traffic flow statistic distribution for reducing the effect of distribution imbalance and overfitting problem during network learning. Besides, the deep bidirectional long short-term memory network (DBLSTM) is introduced to reduce the uncertainty problem by forward and reverse contexts approximation process in the stochasticity reducing layer, and then the long short-term memory network (LSTM) is used to forecast the next traffic flow state in the forecasting layer. Furthermore, sufficient comparative experiments have been conducted and the results show the proposed model has better ability on solving uncertainty and overfitting problems than the state-of-art methods.

Point of Interest (POI) is an important task in Location Based Social Networks like, Gowalla, Foursquare and Brightkite etc as it guides user in decision making process. Although, ‘‘Probabilistic Factor Model’’ (PFM), is a promising method applied in various recommender systems, it is not effective in sparse data situation. Moreover, traditional recommendation methods do not consider the social feature of consumption, where a reliable friend suggests a popular item that does not pair with our preferences. This paper contributes a state-of-the-art solution for a personalized recommendation tool, which suggests interesting and new locations to users by bridging preference-aware and social-based recommendations. To demonstrate the effect of social context of user, the social regularization term, is unified into the Poisson Factor Model. Experiments are performed using the Gowalla and Brightkite dataset. The experiment reveals that, the new framework achieves higher accuracy than the traditional recommendation algorithm.