Search

The widespread communication of academic ideas and research achievements through literature and media has generated massive academic big data. Analyzing such academic big data and discovering knowledge can discover comprehensive and predictive academic intelligence and provide corresponding services, which are valuable for scholars, journals, institutions and governments for career assessment, topic selection, funding management and resource allocation. This paper proposes a heterogeneous knowledge-learning method for understanding and predicting academic impact in the transportation field. The proposed method is illustrated on the academic big data collected from the papers published in 34 transportation journals from 2008 to 2018. We extract four types of features including bibliometric, altmetric, network and semantic features, and build hybrid feature embedding via TransR and Doc2vec that involving domain knowledge. Further, an academic impact prediction model for articles named as Hy-LSTM-Att is proposed, which weighs the hybrid features by the attention mechanism and predicts academic impact with the bi-LSTM recurrent neural networks. Experimental results demonstrate that the proposed Hy-LSTM-Att model outperforms competing shallow and deep learning models. Additionally, the feature ablation experiments illustrate that the four types of features positively contribute to the performance of impact prediction.

With the continuous expansion of aquaculture scale and density, contemporary aquaculture methods have been forced to overproduce resulting in the accelerated imbalance rate of water environment, the frequent occurrence of fish diseases, and the decline of aquatic product quality. Moreover, due to the fact that the average age profile of agricultural workers in many parts of the world are on the higher side, fishery production will face the dilemma of shortage of labor, and aquaculture methods are in urgent need of change. Modern information technology has gradually penetrated into various fields of agriculture, and the concept of intelligent fish farm has also begun to take shape. The intelligent fish farm tries to deal with the precise work of increasing oxygen, optimizing feeding, reducing disease incidences, and accurately harvesting through the idea of “replacing human with machine,” so as to liberate the manpower completely and realize the green and sustainable aquaculture. This paper reviews the application of fishery intelligent equipment, IoT, edge computing, 5G, and artificial intelligence algorithms in modern aquaculture, and analyzes the existing problems and future development prospects. Meanwhile, based on different business requirements, the design frameworks for key functional modules in the construction of intelligent fish farm are proposed.

Background Thallium poisoning is rare and difficult to recognize. Early diagnosis and treatment of thallium-poisoned patients are essential to prevent morbidity and mortality. Aim To evaluate the efficacy of treatments and outcomes of five patients with early diagnosis of acute thallium poisoning. Methods Five patients who consumed a thallium-contaminated meal were hospitalized in succession, and underwent clinical examinations such as blood tests and electromyography tests. Urine and blood tests confirmed the diagnosis of thallotoxicosis, revealing the occurrence of food poisoning. All patients underwent detoxification treatment, including hemoperfusion (HP) and treatment with Prussian blue (PB). A 24-mo follow-up was performed to evaluate the long-term outcomes on the patients after discharge. Results Initially, the patients presented with symptoms of acute thallium poisoning including hyperalgesia of the limbs and abdominalgia, which may differ from common peripheral neuropathy. Accompanying symptoms such as hepatic damage and alopecia were observed in all the patients, which further confirmed the diagnosis of poisoning. Treatment with chelating agents was ineffective, while HP and treatment with PB drastically decreased the thallium concentration in the urine and blood. With early diagnosis and intervention, four patients had a good prognosis and no permanent sequelae. One patient developed blindness and disability during the 24-mo follow-up period. Conclusion Identification of incident cluster and characteristic symptoms is extremely important for early diagnosis of acute thallium poisoning. HP plus PB is essential to improve the prognosis of thallium-poisoned patients.

Anti-α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor encephalitis is an anti-neuronal surface antigen autoimmune encephalitis and is relatively rare. Our study evaluated a patient who developed anti-AMPA2 receptor encephalitis with memory deficits and refractory focal seizures as paroxysmal jerking on right face as well as dystonic seizure on right hand. On this patient, the combination treatment of levetiracetam, carbamazepine, and clonazepam, monthly periodic intravenous immunoglobin and immunosuppressive therapies for 5 months was not effective for the focal seizures, while his memory loss was slightly improved. However, adjunctive perampanel treatment led to a rapid relief of seizures. Perampanel is suggested in seizures associated with anti-AMPA receptor encephalitis by directly attenuating nerve hyperexcitability caused by glutamate and Ca2+-permeable GluA4 subunit of AMPA receptors.

We develop a dynamic model to study the systemic risk of the banking network, so as to study the dynamics of bank defaults. In contrast to the existing literature, we show that while the possibility of contagion is determined by interconnectedness of the financial network, whether a financial crisis can occur depends on the profile of the liquid assets of the banks in the system. Based on the dynamic model, we introduce a time to crisis index that allows us to predict the occurrence of a financial crisis. We then provide an intuitive measure of systemic risk. To illustrate the potential usefulness of our model, we provide an analysis of the system of twenty‐two German banks. We show how many of the banks are fundamentally weak, where the contagion effect may arise from, how strong the contagion effect is, and how significant the systemic risk is.

Atrial fibrillation (AF) is a major cardiovascular disease that has affected thousands of individuals worldwide. The electrocardiogram (ECG) is the most extensively applied approach to detect AF at present, while the traditional detection strategy based on the visual observation of ECG data is often laborious and inefficient. In this work, we specially designed an intelligent recognition system based on a novel convolutional neural network that utilizes the multi-scale convolution kernel and bidirectional gated recurrent unit with attention mechanism for AF detection. Also, two standard control groups using 10-fold cross-validation were performed to assess the validity of the proposed model. The empirical results not only demonstrate the high efficiency of multi-scale convolution kernel, but also show that the model has a more superior classification performance to several state-of-the-art methods with an accuracy of 98.3% and 97.7% on two public databases, respectively. Due to its high performance, we plan to develop the model into portable devices to benefit more individuals such as the elderly and athletes.

In this article, we address the problem of visual question generation (VQG), a challenge in which a computer is required to generate meaningful questions about an image targeting a given answer. The existing approaches typically treat the VQG task as a reversed visual question answer (VQA) task, requiring the exhaustive match among all the image regions and the given answer. To reduce the complexity, we propose an innovative answer-centric approach termed radial graph convolutional network (Radial-GCN) to focus on the relevant image regions only. Our Radial-GCN method can quickly find the core answer area in an image by matching the latent answer with the semantic labels learned from all image regions. Then, a novel sparse graph of the radial structure is naturally built to capture the associations between the core node (i.e., answer area) and peripheral nodes (i.e., other areas); the graphic attention is subsequently adopted to steer the convolutional propagation toward potentially more relevant nodes for final question generation. Extensive experiments on three benchmark data sets show the superiority of our approach compared with the reference methods. Even in the unexplored challenging zero-shot VQA task, the synthesized questions by our method remarkably boost the performance of several state-of-the-art VQA methods from 0% to over 40%. The implementation code of our proposed method and the successfully generated questions are available at https://github.com/Wangt-CN/VQG-GCN .

Brucella endocarditis is a fatal complication and the most frequent cause of death for human brucellosis. This study aimed to systematically review the literature on the follow-up outcomes of Brucella endocarditis and analyze the determinants affecting the follow-up outcomes. The databases PubMed, Web of Science, Embase, and Cochrane were searched using keywords and suitable combinations. All studies reporting the follow-up outcomes of Brucella endocarditis were included. Finally, a total of 76 studies (207 patients), including cases or case series, were included. The event rate for patients who underwent short- and long-term follow-up was 12.0% (2 relapsed and 1 died) and 8.1% (6 relapsed and 8 died), respectively. The differences in outcomes between different age groups (18–39, 40–59, and ≥60) were significant (P

To ensure the safe and effective application of deep geotechnical engineering, it is necessary to establish a suitable constitutive model of the rocks in thermal environment, including geothermal energy mining, and deep geological treatment of nuclear waste and tunnel fire. A strain-softening damage model of the rocks with defect growth is established based on damage evolution. According to the change in the tangent modulus from the compression phase to the elastic stage, we established a constitutive model reflecting the nonlinearity characteristics of the compression phase. According to the change in the energy conversion, the energy consumption coefficient is introduced to describe the deformation characteristics of the rocks when the residual stress is reached. The specific physical meaning and the calculation method of the energy consumption coefficient are given to correct the shortcomings of the evolution of the damage variable. We then analyzed the relationship between parameters in the constitutive model of the rocks at high temperature and normal temperature. Based on the constitutive model of the rocks at normal temperature, we established a constitutive model reflecting the influence of high temperature. Finally, the established model fit the stress-strain curves of several kinds of the rocks at different temperatures and generated desired results. The established model can be used to describe the stress-strain curve characteristics of the rocks at different temperatures.

Feature selection, which eliminates irrelevant and redundant features, is one of the most efficient classification methods. However, searching for an optimal subset from the original set is still a challenging problem. This paper proposes a novel feature selection algorithm named hybrid improved dragonfly algorithm (HIDA) which combines the advantages of both mRMR and improved dragonfly algorithm (IDA) in order to generate promising candidate subset and achieve higher classification accuracy rate. Firstly, to generate promising subset, features with small weight have chance to be selected into candidate subset with a small probability in mRMR. Secondly, to balance the exploitation and exploration capabilities of IDA, dynamic swarming factors are proposed to balance global and local capability. Lastly, to enhance the exploitation capability of IDA, quantum local optimum and global optimum are introduced in the position updating mechanism. The performance of HIDA is investigated on ten gene expression datasets and eight UCI data sets from the UCI Machine Learning Data Repository. Results show that the performance of HIDA is superior to BBA, BDA, CDA, LBPSO, MPMDWOA and MSMCCS.

The composition of the intestinal flora of patients with Parkinson's disease (PD) can change. However, whether reshaping the gut microbial composition can treat PD remains to be seen. The present study evaluated the effect of intestinal flora in the treatment of PD in a C57BL/6 mouse PD model induced by 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP). Chronic, low-dose, MPTP-treated mice exhibited upregulated gene expression levels of TNF-α and IL-1β in the substantia nigra (SN) of the mice, and induced intestinal microbial disorders. This indicated that the chronic low-dose MPTP model could be used to evaluate the progress of early intestinal pathology and intestinal flora imbalance in PD. After transplantation of faecal bacteria to MPTP-induced PD mice, the level of inflammation in the SN of the mice was reduced, and motor dysfunction was alleviated. Notably, faecal microbiota transplantation (FMT) upregulated the abundance of Blautia but downregulated Anaerostipes, Bifidobacterium, ASF356 and Ruminococcus in the gut of PD mice. In addition, FMT reduced the activation of microglia and astrocytes in the SN and reduced the expression levels of GSK3β, IL-1β, inducible nitric oxide synthase and phosphorylated PTEN in the SN. Overall, the present study demonstrated that gut microbial dysfunction is associated with the pathogenesis of PD, and that FMT can protect PD mice by inhibiting neuroinflammation.

Mental health prediction is one of the most essential parts of reducing the probability of serious mental illness. Meanwhile, mental health prediction can provide a theoretical basis for public health department to work out psychological intervention plans for medical workers. The purpose of this paper is to predict mental health of medical workers based on machine learning by 32 factors. We collected the 32 factors of 5,108 Chinese medical workers through questionnaire survey, and the results of Self-reporting Inventory was applied to characterize mental health. In this study, we propose a novel prediction model based on optimization algorithm and neural network, which can select and rank the most important factors that affect mental health of medical workers. Besides, we use stepwise logistic regression, binary bat algorithm, hybrid improved dragonfly algorithm and the proposed prediction model to predict mental health of medical workers. The results show that the prediction accuracy of the proposed model is 92.55%, which is better than the existing algorithms. This method can be used to predict mental health of global medical worker. In addition, the method proposed in this paper can also play a role in the appropriate work plan for medical worker.

Respiratory diseases, especially the pandemic of respiratory infectious diseases and refractory chronic lung diseases, remain a key clinical issue and research hot spot due to their high prevalence rates and poor prognosis. In this review, we aimed to summarize the recent advances in the therapeutic effects and molecular mechanisms of key common bioactive compounds from Chinese herbal medicine. Based on the theories of traditional Chinese medicine related to lung diseases, we searched several electronic databases to determine the high-frequency Chinese medicines in clinical application. The active compounds and metabolites from the selected medicines were identified using the Traditional Chinese Medicine Systems Pharmacology Database (TCMSP) by analyzing oral bioavailability and drug similarity index. Then, the pharmacological effects and molecular mechanisms of the selected bioactive compounds in the viral and bacterial infections, inflammation, acute lung injury (ALI), chronic obstructive pulmonary disease (COPD), pulmonary fibrosis, asthma, and lung cancer were summarized. We found that 31 bioactive compounds from the selected 10 common Chinese herbs, such as epigallocatechin-3-gallate (EGCG), kaempferol, isorhamnetin, quercetin, and β-sitosterol, can mainly regulate NF-κB, Nrf2/HO-1, NLRP3, TGF-β/Smad, MAPK, and PI3K/Akt/mTOR pathways to inhibit infection, inflammation, extracellular matrix deposition, and tumor growth in a series of lung-related diseases. This review provides novel perspectives on the preclinical study and clinical application of Chinese herbal medicines and their bioactive compounds against respiratory diseases.

In this paper, we propose to investigate out-of-domain visio-linguistic pretraining, where the pretraining data distribution differs from that of downstream data on which the pretrained model will be fine-tuned. Existing methods for this problem are purely likelihood-based, leading to the spurious correlations and hurt the generalization ability when transferred to out-of-domain downstream tasks. By spurious correlation, we mean that the conditional probability of one token (object or word) given another one can be high (due to the dataset biases) without robust (causal) relationships between them. To mitigate such dataset biases, we propose a Deconfounded Visio-Linguistic Bert framework, abbreviated as DeVLBert1, to perform intervention-based learning. We borrow the idea of the backdoor adjustment from the research field of causality and propose several neural-network based architectures for Bert-style out-of-domain pretraining. The quantitative results on three downstream tasks, Image Retrieval (IR), Zero-shot IR, and Visual Question Answering, show the effectiveness of De-VLBert by boosting generalization ability 2.

As global scientific researches are witnessing rapid and high-quality development, cooperation has become the development trend in scientific research. This article models social networks, explains in detail two common propagation models, and gives the definition of the problem of maximizing influence in social networks. On the basis of which, the maximization of influence is introduced into the scientific collaboration network. It’s mainly used to describe the cooperative contribution between scientific research scholars, the scientific research innovation ability and the academic leadership ability of the cooperative team. This paper proposes a method to find authors with greatest influence in scientific collaboration network, in order to mine the deep information within the scientific collaboration network.

BACKGROUND: To identify the factors related to the effectiveness of retrograde removal of double J ureteral stents using a simple snare technique in female patients. METHODS: It was a retrospective observational study. From 2012 to 2017, 128 female patients underwent fluoroscopy-guided, retrograde double J stent removal with or without replacement using a simple snare technique. The position of the tip of the double J stent in the urinary bladder was classified according to position types A, B, C, and D, corresponding to the ipsilateral lateral, ipsilateral medial, contralateral medial, and contralateral lateral quadrants of the urinary bladder, respectively. The factors influencing the fluoroscopy time upon removal of the double J stent were analyzed. RESULTS: Technical success rate of 312 procedures performed was 97.44% (304/312) with the mean fluoroscopy time of 12.68±7.34 minutes. Eight double J stent removal procedures failed, and all of these occurred in the position type D patients. The position type and the number of curls of the double J stent had a significant influence on the fluoroscopy time (P

Objective The aim of the study is to compare the efficacy and safety of angioplasty alone with acute stenting for acute tandem occlusions (TO) due to internal carotid artery atherosclerotic. Methods We identified 112 patients who underwent an endovascular treatment (EVT) for acute tandem internal carotid artery occlusions from the prospectively maintained registries 5 comprehensive stroke centers. The study cohort included 75 patients with underlying atherosclerotic lesion of the extracranial internal carotid artery, forty-five in the balloon angioplasty (BA) alone group and 30 in the acute stenting (AS) group. Using propensity score matching analysis, forty-four patients were matched. Clinical characteristics and outcome data were compared between two groups. Results The successful reperfusion immediately post procedure [72.7% (16/22) vs. 77.3% (17/22), P = 1.0] and 90-days good functional outcome [54.5% (12/22) vs. 59.1% (13/22), P = 0.761] were not significantly different between the BA group and AS group. There was also no significant difference in the rate of symptomatic intracranial hemorrhage [13.6% (3/22) vs. 9.1% (2/22), P = 1.00] and restenosis of ICA (>50%)[27.3% (6/22) vs. 22.7% (5/22), P = 0.728] between 2 groups. Patients in the BA group appear to have a numerically lower rate of asymptomatic intracranial hemorrhage [40.9% (9/22) vs. 50% (11/22), P = 0.545] and mortality [0 vs. 9.1% (2/22), P = 0.488] than in the AS group, although there were not statistically significant. Conclusion Among TO patients with etiology of large vessel atherosclerosis, no statistical differences in outcome are identified between balloon angioplasty alone versus acute stenting. Future randomized controlled trials are warranted.

Grouted sleeve connections have been widely used in precast construction for joining rebar. In order to study their post-fire seismic performance, grouted sleeve connections, which are heated first to elevated temperature and then cooled down to room temperature, are tested under both monotonic and cyclic loadings. For comparisons, the same tests are also carried out on continuous rebar. The failure modes, skeleton curves and compression strength of the specimens are analyzed to study the influence of temperature on the seismic performance of the grouted sleeve connections. The results show that buckling occurs to all the grouted sleeve connections under cyclic loading. It is also found that the skeleton curves, compressive and buckling strength of both the connections and the continuous rebar decrease with the increase of temperature up to 600 °C. From the test results and on the basis of the GMP and GA hysteric constitutive models of steel, a modified model for cyclic constitutive law of the continuous rebar and the grouted sleeve connections is presented and validated. The results demonstrate that the modified model is capable of accurately simulating hysteretic properties of steel rebars and grouted sleeve connections.

Although repetitive transcranial magnetic stimulation (rTMS) has been considered a potentially effective treatment for cognitive impairment in patients with Alzheimer's disease (AD), previous studies have produced inconsistent results. The objective of this meta-analysis was to evaluate the effects of rTMS on cognitive function in patients with AD.PubMed, EMBASE, Web of Science, MEDLINE and Cochrane Central Register of Controlled Trials (CENTRAL) were searched for relevant terms. Abstracts of all papers were carefully reviewed, followed by data extraction, quality assessment, data synthesis and subgroup analyses.A total of 12 studies with 231 patients were included, with 8 randomized controlled studies and 4 self-controlled studies. Eleven studies used high frequency rTMS (≥ 5 Hz), but only one study directly compared the difference between low-frequency (1 Hz) and high-frequency (20 Hz). Random-effects analysis revealed that rTMS could significantly improve cognition compared with sham-rTMS (SMD: 0.60, 95% CI: 0.35-0.85, P .0001). In subgroup analyses, the effect for stimulation at a single target was 0.13 (95% CI: -0.35-0.62) and multiple targets 0.86 (95% CI: 0.18-1.54). Treatment for ≤3 sessions produced an effect of 0.29 (95% CI: -1.04-1.62), whereas treatment for ≥5 sessions produced an effect of 2.77 (95% CI: 2.22-3.32). No differences were found for rTMS combined with medication or cognitive training.rTMS can significantly improve cognitive ability in patients with mild to moderate AD. Stimulation of multiple sites and long-term treatment are better at improving AD-associated cognitive performance. Furthermore, some novel interventional targets, like precuneus (PC), may be a more effective therapeutic site to improve memory in AD.

We present a novel counterfactual framework for both Zero-Shot Learning (ZSL) and Open-Set Recognition (OSR), whose common challenge is generalizing to the unseen-classes by only training on the seen-classes. Our idea stems from the observation that the generated samples for unseen-classes are often out of the true distribution, which causes severe recognition rate imbalance between the seen-class (high) and unseen-class (low). We show that the key reason is that the generation is not Counterfactual Faithful, and thus we propose a faithful one, whose generation is from the sample-specific counterfactual question: What would the sample look like, if we set its class attribute to a certain class, while keeping its sample attribute unchanged? Thanks to the faithfulness, we can apply the Consistency Rule to perform unseen/seen binary classification, by asking: Would its counterfactual still look like itself? If ``yes'', the sample is from a certain class, and ``no'' otherwise. Through extensive experiments on ZSL and OSR, we demonstrate that our framework effectively mitigates the seen/unseen imbalance and hence significantly improves the overall performance. Note that this framework is orthogonal to existing methods, thus, it can serve as a new baseline to evaluate how ZSL/OSR models generalize. Codes are available at https://github.com/yue-zhongqi/gcm-cf., Accepted by CVPR 2021

Recent years have witnessed enormous progress of online learning. However, a major challenge on the road to artificial agents is concept drift, that is, the data probability distribution would change where the data instance arrives sequentially in a stream fashion, which would lead to catastrophic forgetting and degrade the performance of the model. In this paper, we proposed a new Bilevel Online Deep Learning (BODL) framework, which combine bilevel optimization strategy and online ensemble classifier. In BODL algorithm, we use an ensemble classifier, which use the output of different hidden layers in deep neural network to build multiple base classifiers, the important weights of the base classifiers are updated according to exponential gradient descent method in an online manner. Besides, we apply the similar constraint to overcome the convergence problem of online ensemble framework. Then an effective concept drift detection mechanism utilizing the error rate of classifier is designed to monitor the change of the data probability distribution. When the concept drift is detected, our BODL algorithm can adaptively update the model parameters via bilevel optimization and then circumvent the large drift and encourage positive transfer. Finally, the extensive experiments and ablation studies are conducted on various datasets and the competitive numerical results illustrate that our BODL algorithm is a promising approach.

As the main factor of China’s higher education, effective evaluation of the scientific research efficiency of domestic universities is extremely significant to improve the scientific research management system of local universities, optimize the allocation of scientific research resources, and promote regional economic development. Based on the ten-year statistical data of the compilation of science and technology statistics of colleges and universities, this paper adopts the data envelopment analysis (DEA) model to study the scientific research efficiency of 31 provincial universities in Shandong Province from three dimensions: level, type, and region. The results show that, first, the overall scientific research efficiency of the sample universities is not high and takes on a downward trend; there are some differences, changes in the same direction, and level convergence in each dimension of scientific research efficiency. Second, there is a positive correlation between the comprehensive scientific research strength of the sample universities and the technical efficiency. Third, the level of regional economic development has a positive impact on the scale efficiency of scientific research, which is closely related to the technical efficiency, and there is no positive correlation between these two. Thus, the research suggests that, in the future, colleges and universities should formulate differentiated management policies, build a “Government University” two-level scientific research performance evaluation system guided by scientific research quality, change the concept of scientific research managers, and improve the ability of personnel to improve the efficiency of scientific research in colleges and universities.

PURPOSE To compare the outcomes of self-expandable metal stent placement and percutaneous gastrostomy (PG) for the treatment of patients with esophageal cancer (EC) and dysphagia. MATERIALS AND METHODS This retrospective observational study consisted of 113 patients with EC and dysphagia who underwent either stent placement (n = 47) or PG (n = 66) at a single center between June 2014 and June 2018. RESULTS There were 63 men and 50 women, with a mean age of 76.5 years (standard deviation 4.9 years). The 2 groups had similar baseline characteristics, except that the PG group had a higher percentage of patients with cervical EC (22.7% vs 2.1%, P < .001). The PG group had better maintenance of nutritional status in terms of reduction in serum albumin level (P = .039) and weight loss (P = .041). Compared with the stent group, the PG group demonstrated a lower incidence of local severe pain (0% vs 21.3%, P < .001) and lower incidence of dislodgment of device (1.5% vs 19.1%, P = .002). The PG group demonstrated longer overall survival compared with the stent group for Stages II and III (201 vs 185 days, P = .034) and Stage IV (122 vs 86 days, P = .001). CONCLUSIONS Compared with stent insertion, PG is associated with better maintenance of nutritional status, fewer complications, and better survival. Thus, PG may be the preferred choice for treating malnutrition in patients with EC and dysphagia.

Traditional approach of blockchain requests system to generate private key by random algorithm. Then the system generates public key and wallet address through one-way algorithm. User in blockchain only needs to use his private key and public key to finish a transaction. Since this it faces problems that system couldn’t solve the situation that a stranger gets private key of a user, he can do any operation in blockchain as the user. This paper presents Pftom, a trading blockchain system based on fingerprint recognition. Pftom generates private key by randomly selecting a user’s input fingerprint image and creates a local fingerprint feature database to recognize the identity when a sensitive operation is being done. These improved the security of the system and lowered the cost of providing services than centralized transaction systems do. The proposed system was evaluated with two experiments on security and performance. We found Pftom can totally prevent others from logging in a user’s account. It can also hold a large number of concurrent transactions.

In this paper, we propose to investigate the problem of out-of-domain visio-linguistic pretraining, where the pretraining data distribution differs from that of downstream data on which the pretrained model will be fine-tuned. Existing methods for this problem are purely likelihood-based, leading to the spurious correlations and hurt the generalization ability when transferred to out-of-domain downstream tasks. By spurious correlation, we mean that the conditional probability of one token (object or word) given another one can be high (due to the dataset biases) without robust (causal) relationships between them. To mitigate such dataset biases, we propose a Deconfounded Visio-Linguistic Bert framework, abbreviated as DeVLBert, to perform intervention-based learning. We borrow the idea of the backdoor adjustment from the research field of causality and propose several neural-network based architectures for Bert-style out-of-domain pretraining. The quantitative results on three downstream tasks, Image Retrieval (IR), Zero-shot IR, and Visual Question Answering, show the effectiveness of DeVLBert by boosting generalization ability., 10 pages, 4 figures, to appear in ACM MM 2020 proceedings

The survival benefits and which patients with advanced hepatocellular carcinoma (HCC) would benefit from sorafenib plus transarterial chemoembolization (TACE) therapy remain controversial. We aimed to develop a prognostic score model for predicting different prognoses of patients with HCC and portal vein invasion who received sorafenib plus TACE. This observational study included 167 patients with HCC and portal vein invasion undergoing sorafenib combined with TACE from January 2013 to June 2018 at two hospitals. Multivariate Cox regression analysis was performed using a training cohort (n = 83) to identify critical factors associated with survival. Then, a prognostic score model was established to classify different outcomes and confirmed using a validation cohort (n = 84). Three factors were determined to critically impact survival in the training cohort: portal vein invasion extent, sorafenib-related dermatologic response, and initial radiological response. Using the β-coefficients of these factors, a prognostic score was calculated, and the survival time decreased as the score increased. Based on the prognostic score model, three different prognoses of patients with 0 points, 2–3 points, and > 3 points were stratified with a median survival of 38.0 months, 20.0 months, and 7.0 months, respectively (P

The COVID-19 pandemic has caused a global public health crisis. There is a pressing need for evidence-based interventions to address the devastating clinical and public health effects of the COVID-19 pandemic. The Chinese scientists supported by private and government resources have adopted extensive efforts to identify effective drugs against the virus. To date, a large number of clinical trials addressing various aspects of COVID19 have been registered in the Chinese Clinical Trial Registry (ChiCTR), including more than 200 interventional studies. Under such an urgent circumstance, the scope and quality of these clinical studies vary significantly. Hence, this review aims to make a comprehensive analysis on the profiles of COVID-19 clinical trials registered in the ChiCTR, including a wide range of characteristics. Our findings will provide a useful summary on these clinical studies since most of these studies will encounter major challenges from the design to completion. It will be a long road for the outcomes of these studies to be published and international collaboration will help the ultimate goals of developing new vaccines and anti-viral drugs.

We present a novel unsupervised feature representation learning method, Visual Commonsense Region-based Convolutional Neural Network (VC R-CNN), to serve as an improved visual region encoder for high-level tasks such as captioning and VQA. Given a set of detected object regions in an image (e.g., using Faster R-CNN), like any other unsupervised feature learning methods (e.g., word2vec), the proxy training objective of VC R-CNN is to predict the contextual objects of a region. However, they are fundamentally different: the prediction of VC R-CNN is by using causal intervention: P(Y|do(X)), while others are by using the conventional likelihood: P(Y|X). This is also the core reason why VC R-CNN can learn "sense-making" knowledge like chair can be sat -- while not just "common" co-occurrences such as chair is likely to exist if table is observed. We extensively apply VC R-CNN features in prevailing models of three popular tasks: Image Captioning, VQA, and VCR, and observe consistent performance boosts across them, achieving many new state-of-the-arts. Code and feature are available at https://github.com/Wangt-CN/VC-R-CNN., Comment: Accepted by CVPR 2020

We present a novel unsupervised feature representation learning method, Visual Commonsense Region-based Convolutional Neural Network (VC R-CNN 1 ), to serve as an improved visual region encoder for high-level tasks such as captioning and VQA. Given a set of detected object regions in an image (e.g., using Faster R-CNN), like any other unsupervised feature learning methods (e.g., word2vec), the proxy training objective of VC R-CNN is to predict the contextual objects of a region. However, they are fundamentally different: the prediction of VC R-CNN is by using causal intervention: P(Y|do(X)), while others are by using the conventional likelihood: P(Y|X). We extensively apply VC R-CNN features in prevailing models of two popular tasks: Image Captioning and VQA, and observe consistent performance boosts across all the methods, achieving many new state-of-the-arts 2 .

Encryption can protect data in outsourced databases -- in the cloud -- while still enabling query processing over the encrypted data. However, the processing leaks information about the data specific to the type of query, e.g., aggregation queries. Aggregation over user-defined groups using SQL's GROUP BY clause is extensively used in data analytics, e.g., to calculate the total number of visitors each month or the average salary in each department. The information leaked, e.g., the access pattern to a group, may reveal the group's frequency enabling simple, yet detrimental leakage-abuse attacks. In this work we present SAGMA -- an encryption scheme for performing secure aggregation grouped by multiple attributes. The querier can choose any combination of one or multiple attributes in the GROUP BY clause among the set of all grouping attributes. The encryption scheme only stores semantically secure ciphertexts at the cloud and query processing hides the access pattern, i.e., the frequency of each group. We implemented our scheme and our evaluation results underpin its practical feasibility.

A suitable model for OLED simulation is provided. We included the tail state model, field dependent model, and triplet conversion model at heterojunction interface into our developed simulator to provide academic free TCAD simulator for research society. The details of simulator, algorithm, and fitting to experimental results will be provided.

The task of image–text matching refers to measuring the visual-semantic similarity between an image and a sentence. Recently, the fine-grained matching methods that explore the local alignment between the image regions and the sentence words have shown advance in inferring the image–text correspondence by aggregating pairwise region-word similarity. However, the local alignment is hard to achieve as some important image regions may be inaccurately detected or even missing. Meanwhile, some words with high-level semantics cannot be strictly corresponding to a single-image region. To tackle these problems, we address the importance of exploiting the global semantic consistence between image regions and sentence words as complementary for the local alignment. In this article, we propose a novel hybrid matching approach named Cross-modal Attention with Semantic Consistency (CASC) for image–text matching. The proposed CASC is a joint framework that performs cross-modal attention for local alignment and multilabel prediction for global semantic consistence. It directly extracts semantic labels from available sentence corpus without additional labor cost, which further provides a global similarity constraint for the aggregated region-word similarity obtained by the local alignment. Extensive experiments on Flickr30k and Microsoft COCO (MSCOCO) data sets demonstrate the effectiveness of the proposed CASC on preserving global semantic consistence along with the local alignment and further show its superior image–text matching performance compared with more than 15 state-of-the-art methods.

IntroductionEssential tremor (ET), a tremor disorder, is one of the most common movement disorders. Only oral drugs (propranolol, primidone, topiramate, etc）are still the first-line treatment recommended by the Food and Drug Administration. Propranolol is thought to potentially reduce upper limb action tremor. However, it has a poor effect on axial tremor symptoms, such as essential head tremor and voice tremor. Studies have shown that tremor severity develops over time, possibly producing other clinical tremors and neurological soft signs (such as memory loss, gait abnormalities, balance disorders, etc), which further increases the difficulty of treating tremors. However, some recent studies provide emerging evidence for oral propranolol on subgroups of ET, which is based on the anatomical distribution of ET (lower extremities, head, sound, tongue, etc). This systematic review aims to synthesise these new data to improve the efficacy of propranolol in ET subgroups.Methods and analysisWe will search for randomised controlled trials from the PubMed, MEDLINE, EMBASE, Cochrane Library, UptoDate and PEDro databases from inception to June 2019. All data will be extracted independently by two reviewers and compared at the end of the review. The two reviewers will screen the study quality, and the Cochrane Collaboration’s tool in Review Manager (RevMan) V.5.3.3 will be used to evaluate risk of bias. Our primary outcome will be the functional disability component related to tremors, as measured by the Fahn-Tolosa-Marin Tremor Rating Scale subscales B and C. Secondary outcomes will include severity of tremors and quality of life. Narrative and meta-analytical syntheses are planned.Ethics and disseminationPublished aggregated data will be used in this review analysis and therefore no ethical approval is required. The results will be published in peer-reviewed journals, and proliferation activities will include diverse social stakeholders, non-academic groups and patients.PROSPERO registration numberCRD42018112580

With the rapid development of the construction industry, the country has a higher demand for scaffolding engineering, so it is very necessary to develop and promote the application of wheel buckle scaffolding. Steel tube scaffold with wheel buckle has the characteristics of clear transmission and good mechanical performance. In order to study the structural performance of steel tubular scaffolding with wheel buckle, the single span three-step element frame was tested. The failure mode and ultimate bearing capacity of the frame are obtained. The finite element software Sap2000 was used to conduct 3d modeling and linear buckling analysis of scaffolds in the test. The results of experiments and finite element analysis show that the failure type of steel tubular scaffolding is the overall torsional instability failure. The connection stiffness at the joint of the diagonal brace fastener has a great influence on the wheel-buckle scaffold. The diagonal brace has obvious influence on the bearing capacity of steel tubular scaffolding body with buckles.

Grouted sleeve connections are widely used in precast concrete structures for joining rebar. In order to understand the behavior of grouted sleeves under and after fire, tensile tests are carried out on grouted sleeves with threaded bars under three different thermo-mechanical loading conditions, i.e., loading when temperature is constant, heating when a constant mechanical load is maintained and loading at room temperature on previously heated sleeves. Tensile tests of continuous rebar are also conducted for comparisons. Failure modes and degradation of mechanical properties of both grouted sleeve connections and continuous rebar subjected to above conditions are studied. It is found that the temperature resistances of the grouted sleeve connections with or without considering heating process are significantly different, while such differences are negligible for continuous rebar. It is also found that grouted sleeve connections can restore some of the mechanical capacities after they are cooled down to the room temperature. Finally, predictive models are proposed for potential adoption in fire resistance design and post fire evaluation of grouted sleeve connections.

Visual servoing is a hot topic in the field of Robotics. Compared with other sensors, visual servo has the characteristics of high flexibility, high precision and strong robustness of calibration. In this paper, visual servoing are classified according to different types, and the research progress of visual servoing is reviewed from position-based visual servoing, image-based visual servoing, and hybrid visual servoing. Secondly, the feature extraction in the visual servoing process is introduced, and the current development situation is introduced from several feature extraction algorithms Then, the trajectory planning based on image interpolation, the trajectory planning based on optimization and the trajectory planning based on potential field are analyzed and discussed. Finally, based on the current research progress, the future research direction is prospected.

In this research, we processed the GPS and meteorological data from about 220 stations of Crustal Movement Observation Network of China (CMONOC) observed in 2014 and derived the Zenith Total Delay (ZTD) map in both spatial and temporal dimension. The results of ZTD have high accurate and reliable as IGS and all sites with varying locations show obvious variety characteristics of Chinese mainland. Meanwhile, the precipitable water vapor (PWV) correlation coefficients between GPS observation and upper air sounding is close to 1, and the comparison of GPS-derived PWV and observed PWV from meteorological sites indicating GPS observation data generated in CMONOC project applied to the weather forecast research is feasible. In addition, based on all stations covered the whole Chinese land area and using interpolation algorithms, we make contour plots of PWV distribution per hour. We observe obvious feature that the precipitable water in north and western area is less than south and east area all over this year. High latitudes area may be dry and low latitudes area is wet. Keywords: GPS, CMONOC, Precipitable water vapor, Troposphere, ZTD

Background Gastric cancer (GC) is histologically a very heterogeneous disease, and the temporal development of different histological phenotypes remains unclear. Recent studies in lung and ovarian cancer suggest that KRAS activation (KRASact) can influence histological phenotype. KRASact likely results from KRAS mutation (KRASmut) or KRAS amplification (KRASamp). The aim of the study was to investigate whether KRASmut and/or KRASamp are related to the histological phenotype in GC. Methods Digitized haematoxylin/eosin-stained slides from 1282 GC resection specimens were classified according to Japanese Gastric Cancer Association (JGCA) and the Lauren classification by at least two observers. The relationship between KRAS status, predominant histological phenotype and clinicopathological variables was assessed. Results KRASmut and KRASamp were found in 68 (5%) and 47 (7%) GCs, respectively. Within the KRASmut and KRASamp cases, the most frequent GC histological phenotype was moderately differentiated tubular 2 (tub2) type (KRASmut: n = 27, 40%; KRASamp: n = 21, 46%) or intestinal type (KRASmut: n = 41, 61%; KRASamp: n = 23, 50%). Comparing individual histological subtypes, mucinous carcinoma displayed the highest frequency of KRASmut (JGCA: n = 6, 12%, p = 0.012; Lauren: n = 6, 12%, p = 0.013), and KRASamp was more frequently found in poorly differentiated solid type (n = 12, 10%, p = 0.267) or indeterminate type (n = 12, 10%, p = 0.480) GC. 724 GCs (57%) had intratumour morphological heterogeneity. Conclusions This is the largest GC study investigating KRAS status and histological phenotype. We identified a relationship between KRASmut and mucinous phenotype. The high level of intratumour morphological heterogeneity could reflect KRASmut heterogeneity, which may explain the failure of anti-EGFR therapy in GC. Electronic supplementary material The online version of this article (10.1007/s10120-019-00972-6) contains supplementary material, which is available to authorized users.

Rigid bracing dome is a new pre-stressed space structure. In the study, according to the features of the new space structure, a four-step tensioning construction method is proposed and corresponding tensioning control principle is also given. The rigid bracing dome designed in the experimental study has a diameter of 6 m and is equally divided into eight parts in the circumferential direction. The four-step tensioning construction method is explored in detail. The synchronous symmetric tensioning schemes in eight directions, four directions and two directions are studied respectively. The deviations of the experimental model are analyzed by using the orthogonal test design method. The key experimental problems are analyzed to propose corresponding solutions. The construction method is simple and efficient. An appropriate symmetric tensioning scheme can be chosen according to the requirements of practical projects. The deviation of outer ridge cable force is mainly caused by the following factors: vertical coordinates of support and rods, and the lengths of outer ridge cable, inner ridge cable, inner ring rod, and outer ring rod. The coordinate deviations of center rods are mainly affected by the lengths of outer ring rod and inner ridge cable. The construction method and corresponding solutions can be used to guide engineering practices.

Aim We aimed to clarify the characteristics of malignancies in older adults focusing on distant metastasis in the whole body. Methods We retrospectively evaluated 7710 cases of autopsies (4011 men, 3699 women, median age of 80 years), and analyzed the characteristics of metastasis of adenocarcinoma, squamous cell carcinoma and urothelial carcinoma in each organ. Results The total number of cases with adenocarcinoma, squamous cell carcinoma or urothelial carcinoma was 2856, and most of them were adenocarcinomas. Among them, 1604 had metastatic lesions, and patients with metastasis were younger than those without metastasis. The major primary organs of adenocarcinoma were the stomach, colon, lung, prostate, gallbladder and pancreas, whereas those for squamous cell carcinoma were the lung, esophagus and uterus. Urothelial carcinoma cases were found in the urinary bladder, kidney and ureter. Metastatic adenocarcinomas mainly originated from the stomach, colon, lung, pancreas and gallbladder. Metastatic squamous cell carcinomas were from the lung, esophagus and uterus, whereas the kidney, bladder and ureter were the primary origins of metastatic urothelial carcinomas. Squamous cell carcinoma showed the highest incidence of metastasis, suggestive of it being of an aggressive phenotype. Furthermore, metastatic ability and the preferred metastatic sites varied among primary organs. Conclusions We revealed an accurate incidence and the characteristics of metastatic cancer in a large-scale autopsy study of older Japanese patients from one institution. Identifying these features might prompt screening for malignancies, and consequently improve quality of life for older adults. Geriatr Gerontol Int 2018; 18: 211-215.

Carbohydrate antigen 19-9 (CA19-9) is a well-known tumor marker for pancreatobiliary cancer, and several studies have shown that an elevated serum CA19-9 level is associated with more aggressive biological behavior in gastric cancer (GC). However, the clinicopathological characteristics of CA19-9-positive GC remain unclear. We herein report an autopsy case of CA19-9-positive GC in an 84-year-old man who was admitted to our hospital because of paralysis and anemia. Autopsy revealed an ulcerative-invasive tumor measuring 72 × 60 mm in the anterior wall of the gastric body. The tumor had invaded beyond the muscularis propria, and metastasized to the lung, liver, and regional lymph nodes. Histologically, the tumor cell had oval nuclei with abundant clear cytoplasm, and tubular and/or papillary features with prominent lymphovascular permeation and perineural invasion, mimicking pancreatobiliary carcinoma. Immunohistochemically, the tumor cells showed diffuse immunopositivity for CA19-9 and carcinoembryonic antigen. According to a review of cases reported in the literature, CA19-9-positive GCs show clinicopathological characteristics such as antral location, ulcerative-infiltrating gross feature, differentiated histology, prominent lymphatic and venous invasion, higher proportion of metastasis, and higher clinical stage. These results suggest that CA19-9-positive GC is pathologically a distinctive type of tumor with aggressive biological behavior.

Fluorescence spectroscopy is frequently used to analyze the concentration of fluorescent materials in solution. However, in conventional fluorescence spectroscopy, the response between the fluorescence intensity and fluorophore concentration is nonlinear at high concentrations due to uncompensated inner-filter effects (IFE). Many methods to resolve this problem have been developed in recent decades. This review introduces the methods used to correct the IFE, including direct correction and parameter correction. Relevant detection parameters, including the materials, matrices, detection limits, detection instruments and relative standard deviations, are tabulated. The advantages and limitations of these correction techniques are also discussed. Finally, the methods used to correct for the IFE are summarized, and future research directions are discussed.

This paper aims to study the effects of different proportion of astragaloside and curcumin on STZ induced Diabetes Mellitus (DM) model of mice, and to select a better proportion of active components. Its ultimate purpose is to lay a basis for the follow-up research on astragaloside-curcumin capsule. Increase-decrease baseline geometric proportion design method and comprehensive performance evaluation utilised to study the effect of different proportion of astragaloside and curcumin on DM mice models, which have an intravenous tail injection of STZ. The proportions of the two components are 10:0, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8, 0:10 respectively. And we will screen out the optimal composition. Blood glycated serum protein (GSP), hepatic glycogen and insulin tested to observe pathological changes in the pancreas. The mice DM model was copied successfully. Compared with the model group, groups treated with the metformin and with different proportions of astragaloside and curcumin help lower the blood glucose levels and GSP levels, increase glycogen stores of model mice by different degrees, and avoid pathological changes of pancreas in the model mice. The ratio of 3:7 was selected as the optimal one, based on the comprehensive performance evaluation method, followed by the ratio of 4:6. The optimal proportion of DM models is 3:7, followed by 4:6. The ratio of total astragaloside and curcumin can lower blood glucose levels, GSP levels, promote the formation of glycogen, and improve the pathological changes of pancreas in the model mice. Keywords: Astragaloside, Curcumin, Baseline geometric proportion design method, STZ

To investigate the effect of Total Alkali in Leonuri Herba (TALH) on rat ear acne model of serum IL-6 level, Thymus and Spleen Tissue Morphology. The rats were divided into the TALH group at high, medium and low doses (200 mg kg −1 , 100 mg·kg −1 , 50 mg·kg −1 ), the tanshinone group (360 mg·kg −1 ), model group and the blank group. All the rats were administrated once a day for 14 days. At the same time on the first day, the rats in the blank group were injected normal saline in the stomach and also had intradermal injection of sterile saline in the auricle. The rats in other groups (model group were lavaged with physiological saline) were injected intradermally auricle spare Staphylococcus epidermidis bacteria liquid. After the injection, the rat ear swelling rate was calculated for 5 consecutive days in terms of rat auricle thickness measurement. 1 h after the last administration, the eyeball blood was removed, centrifuged to separate serum, and measured to be at serum IL-6 level; The tissue sections of thymus and spleen were observed according to morphology criteria. Compared with the blank group, the level of IL-6 in serum of the rats in the model group was remarkably higher, and those in the thymus and spleen groups were remarkably lower, indicating the rat ear acne model is successful. Compared with the model group, the rats in the TALH group can remarkably make serum IL-6 level decrease, causing significant thymus cortical thickening and increased spleen sections, and remarkably increasing the lymphocytes of thymus and spleen. The effect of large dose in the TALH group is the best and increases with the increase of the dose and curative effect. The TALH treatment of acne may be related to the reduction of serum IL-6 levels and the resistance of the atrophy of the thymus and spleen on the rat ear acne model.

Objective: To investigate the pharmacological effects of Leonuri Herba alkaloids (LHA) on prostate hyperplasia in older rats and the effect mechanism. Methods: Remove bilateral testes from BPH model rats, and conduct subcutaneous injection of testosterone and estradiol. At the same time, feed corresponding drugs to the rats by gastric perfusion for 30d. In the first 27d, conduct bladder fistula surgery. Three days after feeding, carry out the detection of the urine flow dynamics. Eyeball blood taking, determination of serum E2 levels, and quickly remove the prostate, thymus gland, spleen, kidney, lung, and bladder. 1/3 prostate homogenate, determine the level of PACP, T, DHT. 1/3 prostate was determined by mRNA expression in bFGF. The remaining 1/3 prostate was observed by light microscopy. Results: LHA could significantly decrease the animal prostate index, level of DHT, T, PACP, and elevate levels of E2 in the serum. It could also significantly reduce the maximum voiding pressure, intercontraction interval, and bladder resting pressure. Conclusion: LHA has good therapeutic effect on prostatic hyperplasia model rats induced by male and female hormone.