Search

Named entity recognition (NER) is an important task in natural language processing. NER is usually used to classify documents, extract information, and translate languages. However, few studies have used NER types to automatically generate questions. In this paper, we proposed a method named NER2QUES to solve the above problem for a low-resource language such as Vietnamese. NER2QUES was the combining pre-trained language model and sequence-to-sequence model. Specifically, we used BERT to detect NERs in a sentence and then applied a sequence-to-sequence model to automatically generate questions that corresponded to NER’s types. We compared the accuracy of the proposed method to PhoBERT and spaCy on the NER task. Also, we used F1, BLEU, ROUGE, and METEOR to measure the effectiveness of this approach with the rules-based method, T5, and BERT on question generation tasks. The experiment results show that the accuracy of our method is more improved than previous methods’ accuracy of 94% on SQuAD, 89% on XQuAD, and 95% on MLQA. This indicates that using NER to automatically generate questions may enrich question answering systems.

In recent years, Question Answering (QA) systems have increasingly become very popular in many sectors. This study aims to use a knowledge graph and deep learning to develop a QA system for tourism in Vietnam. First, the QA system replies to a user's question about a place in Vietnam. Then, the QA describes it in detail such as when the place was discovered, why the place's name was called like that, and so on. Finally, the system recommends some related tourist attractions to users. Meanwhile, deep learning is used to solve a simple natural language answer, and a knowledge graph is used to infer a natural language answering list related to entities in the question. The study experiments on a manual dataset collected from Vietnamese tourism websites. As a result, the QA system combining the two above approaches provides more information than other systems have done before. Besides that, the system gets 0.83 F1, 0.87 precision on the test set.

Understanding factors that influence technology diffusion is central to clinical translation of novel therapies. We characterized the pattern of adoption for laser interstitial thermal therapy (LITT), also known as stereotactic laser ablation (SLA), in neuro-oncology using the National Inpatient Sample (NIS) database. We identified patients age ≥ 18 in the NIS (2012–2018) with a diagnosis of primary or metastatic brain tumor that underwent LITT or craniotomy. We compared characteristics and outcomes for patients that underwent these procedures. LITT utilization increased ~ 400% relative to craniotomy during the study period. Despite this increase, the total number of LITT procedures performed for brain tumor was 2 comorbidities (OR 0.64, CI95 0.51–0.79) or to be older (OR 0.92, CI95 0.86–0.99) and more likely to be female (OR 1.35, CI95 1.08–1.69), Caucasian compared to Black (OR 1.94, CI95 1.12–3.36), and covered by private insurance compared to Medicare or Medicaid (OR 1.38, CI95 1.09–1.74). LITT hospital stays were 50% shorter than craniotomy (IRR 0.52, CI95 0.45–0.61). However, charges related to the procedures were comparable between LITT and craniotomy ($1397 greater for LITT, CI95 $−5790 to $8584). For neuro-oncology indications, LITT utilization increased ~ 400% relative to craniotomy. Relative to craniotomy-treated patients, LITT-treated patients were likelier to be young, female, non-Black race, covered by private insurance, or with

The current BERT-based question answering systems use a question and a contextual text to find the answer. This causes the systems to return wrong answers or nothing if the text contains irrelevant contents with the input question. Besides, the systems haven’t answered yes-no and aggregate questions yet. Besides that, the systems only concentrate on the contents of text regardless of the relationship between entities in the corpus. This systems cannot validate the answer. In this paper, we presented a solution to solve these issues by using the BERT model and the knowledge graph to enhance a question answering system. We combined content-based and linked-based information for knowledge graph representation learning and classified triples into one of three classes such as base class, derived class, or non-existent class. We then used the BERT model to build two classifiers: BERT-based text classification for content information and BERT-based triple classification for link information. The former was able to make a contextual embedding vector for representing triples that were used to classify into the three above classes. The latter generated all path instances from all meta paths of a large heterogeneous information network by running the Motif Search method of Apache Spark on a distributed environment. After creating the path instances, we produced triples from these path instances. We made content-based information by converting triples into natural language text with labels and considered them as a text classification problem. Our proposed solution outperformed other embedding methods with an average accuracy of 92.34% on benchmark datasets and the Motif Finding algorithm with an average executive time improvement of 37% on the distributed environment.

COVID-19 and the government restrictions in place have seriously affected the face-to-face (F2F) mode of delivery in education and higher education has been one of the hardest hit. Universities around the world had to implement the transition from F2F to online on very short notice. In this paper, the author presents a case study that demonstrates the challenges, steps, and adjustments taken to bring a course that has significant components of communication, teamwork, and project management from a full F2F mode to fully online whilst maintaining as much learner-learner and learner-teacher interactions and students engagement as possible. This has been achieved, not without drawbacks and challenges, via synchronous delivery mode and with private channels and breakout room function to promote intra-team and inter-team communications, and teamwork.

Many indices have been developed to assess the impact of scientific publications by investigators, disciplines, and institutions. The h-index has emerged as a leading tool in the assessment of the productivity of authors. Differences in publication and citation opportunity among specialties create inappropriate conclusions when the h-index is used to compare authors across different disciplines. An alternative, the Radicchi index, hf, has been proposed to assess the impact of publications across disciplines. We curated a database of all articles published from 2002 to 2015 from the 3 highest impact factor medical journals: New England Journal of Medicine (NEJM), the Lancet, and the Journal of American Medical Association (JAMA). With this database, articles were categorized into medical subspecialties. We calculated the respective h-index and Radicchi index of each specialty. We found that the Radicchi index eliminated variability associated with publication and citation opportunity between different specialties when compared to the h-index. The Radicchi index is a useful measure of scientific impact and productivity that advances the h-index by allowing interspecialty comparisons. There remains a need to define a researcher's specialty designation especially if he/she conducts multidisciplinary research.

Despite the poor prognosis in patients with type 2 myocardial infarction (MI), no prospective data on risk stratification exists. The aim of this study was to develop and validate a model for prediction of 18-month mortality of among patients with type 2 MI (T2MI) and compare its performance with GRACE and TARRACO scores.The prospective observational study included 712 consecutive patients diagnosed with MI undergoing coronary angiography24 h between January 2017 and December 2018. Diagnosis of T2MI was adjusted according to Third universal definition. A prognostic model was developed by using Bayesian approach and logistic regression analysis with identifying predictors for mortality. The model was validated by bootstrap validation. Comparison performance between scores using Delong test.T2MI was identified in 174 (24.4%) patients. The median age of patients was 69 years, 52% were female. The mortality rate was 20.1% at 18 months. Prior MI, presence of ST elevation, hemoglobin level at admission, Charlson comorbidity index and were independently associated with 18-month mortality. The model to predict 18-month mortality showed excellent discrimination (optimism corrected c-statistic = 0.822) and calibration (corrected slope = 0.893). GRACE and TARRACO scores had moderate discrimination [c-statistic = 0.748 (95% CI 0.652-0.843) and 0.741, 95% CI 0.669-0.805), respectively] and inferior compared with model (p = 0.043 and 0.037, respectively).The risk of mortality among T2MI patients could be accurately predicted by using common clinical characteristics and laboratory tests. Further studies are required with external validation of nomogram prior to clinical implementation.

Injuries from Less-lethal Weapons during the George Floyd Protests In this analysis involving patients treated in Minnesota during protests after George Floyd’s death, 89 patients were identified w...

While atherosclerotic plaque disruption remains the hallmark of type 1 myocardial infarction (T1MI), multiple other mechanisms provoking myocardial supply/demand mismatch (eg, anemia and tachyarrhythmias) are recognized as the potential causes of type 2 myocardial infarction (T2MI). In clinical practice, angiography is underutilized in patients with MI that have typical T2MI triggers, although the presence of these triggers and various forms of atherosclerotic coronary artery disease is not mutually exclusive. We describe a 70-year-old man that developed MI during hospitalization for gastrointestinal bleeding. He was treated conservatively without angiography due to posthemorrhagic anemia, which is a recognized T2MI trigger, and subsequently developed refractory cardiogenic shock. Autopsy revealed atherothrombosis, which is characteristic of T1MI.

A question answering (QA) system based on natural language processing and deep learning is a prominent area and is being researched widely. The Long Short-Term Memory (LSTM) model that is a variety of Recurrent Neural Network (RNN) used to be popular in machine translation, and question answering system. However, that model still has certainly limited capabilities, so a new model named Bidirectional Encoder Representation from Transformer (BERT) emerged to solve these restrictions. BERT has more advanced features than LSTM and shows state-of-the-art results in many tasks, especially in multilingual question answering system over the past few years. Nevertheless, we tried applying multilingual BERT model for a Vietnamese QA system and found that BERT model still has certainly limitation in term of time and precision to return a Vietnamese answer. The purpose of this study is to propose a method that solved above restriction of multilingual BERT and applied for question answering system about tourism in Vietnam. Our method combined BERT and knowledge graph to enhance accurately and find quickly for an answer. We experimented our crafted QA data about Vietnam tourism on three models such as LSTM, BERT fine-tuned multilingual for QA (BERT for QA), and BERT+vnKG. As a result, our model outperformed two previous models in terms of accuracy and time. This research can also be applied to other fields such as finance, e-commerce, and so on.

Rotary machines, and limacon machines in particular, offer a better power to weight ratio compared to reciprocating machines; however, leakage due to improper apex and side sealing have prevented rotary machines from thriving. In this paper, a modelling approach is presented to analyse the vibration of apex seal during the machine operation and the power loss caused by the seal friction. The seal and spring are modelled as a spring-mass system in which the seal deformation is negligible. The seal-groove relative positions have then been categorised into nine different possible cases based on the number of contact points between the seal and the seal groove. A case study has been presented to demonstrate the reliability of the model.

Background: Atherothrombosis is the principal mechanism of type 1 (T1) myocardial infarction (MI), while type 2 (T2) MI is typically diagnosed in the presence of triggers (anemia, arrhythmia, etc.). We aimed to evaluate the proportions of T1 vs. T2 MI based on angiographic and clinical definitions, their concordance and prognosis. Methods: Consecutive MI patients [n = 712, 61% male; age 64.6 ± 12.3 years] undergoing coronary angiography were classified according to the presence of atherothrombosis and identifiable triggers. Association of angiographic and clinical MI type criteria with adverse outcomes (Time follow-up was 1.5 years) was evaluated. Predictive ability of GRACE risk score for all-cause mortality was then assessed. Results: Atherothrombosis and clinical triggers were identified in 397 (55.6%) and 324 (45.5%) subjects, respectively. Only 247 (34.7%) patients had “true” T1MI (atherothrombosis+ / triggers−); 174 (24.4%) were diagnosed with “true” T2MI (atherothrombosis− / triggers+), while 291 (40.9%) had discordant clinical and angiographic characteristics. All-cause mortality in T2MI (20.1%) patients was higher than in T1MI (9.3%), P = 0.002. Presence of triggers [odds ratio (OR) 2.4, 95% CI 1.5-3.6, P < 0.0001] but not atherothrombosis [OR 0.8, 95% confidence interval (CI) 0.5-1.3, P = 0.26] was associated with worse prognosis. GRACE score is a better predictor of death in T1MI vs. T2MI: area under curve 0.893 (95% CI 0.830-0.956) vs 0.748 (95% CI 0.652-0.843), P = 0.013 Conclusion: Angiographic and clinical definitions of MI type are discordant in a substantial proportion of patients. Clinical triggers are associated with all-cause mortality. Predictive performance of GRACE score is worse in T2MI patients.

Ultrasound imaging has found applications in several areas of interventional urology prior to treatment (diagnosis), after treatment (follow-up), and also during treatment (monitoring/guidance). Contrast-enhanced ultrasound (CEUS) imaging and elastographic imaging are the two widely used imaging techniques. CEUS modalities, such as harmonic and multi-pulse imaging, are increasingly applied in interventional urology, because the contrast agents give a clear enhancement of blood vessels in the tissue. Among others, CEUS is used to study the microvasculature of the prostate and kidney for cancer detection to more accurately identify malignant lesions. Elastography is a technique that has been coined in 1991 for the quantification of elastic properties of biological tissue. Various elastographic techniques have been developed, such as Acoustic radiation force imaging (ARFI) and shear wave elastography. These techniques have found various applications in interventional urology such as monitoring the degree of fibrosis in renal allografts after transplantation and the identification of renal masses and prostatic lesions.These enhanced ultrasound modalities represent novel techniques in the evaluation of the prostate and kidney in urology. There are multiple putative benefits of enhanced ultrasound modalities, including improved targeting for prostate biopsy; improved characterization of suspicious renal masses, especially when contrast-enhanced axial imaging (CT or MRI) are contraindicated; and lower cost than CT or MRI. Further investigations and refinements are necessary to define the role of these techniques in current urologic practice.

Limacon is a simple and yet reliable technology, which can be employed to manufacture gas expanders to extract work or electrical power from low-grade heat or solar power resources; Limacon technology can also be applied to small-scale power generation applications or can be used to improve energy efficiency of existing processes. Sultan (Journal of Mechanical Design, 787–793, 2006) and Sultan and Schaller (Journal of Engineering for Gas Turbines and Power, 2011) have come up with an optimum design of Limacon-to-Limacon expanders based on their thermodynamics performance; the design presented a model to calculate the cross-sectional area as well as the velocity of fluid flow through the inlet and discharge ports. However, the problem of optimum geometric characteristics of the inlet and discharge manifolds and the best parameters of various Limacon embodiments are left to be solved. Additionally, the effect of the phase change on the expander performance is yet to be investigated. Aim of this document is to provide the framework of the Limacon machines to achieve optimum expander geometries for power generation systems with various types of working fluids.

To optimize the cost and performance of complex cloud services under dynamic requirements, workflows and diverse cloud offerings, we rely on different elasticity control processes. An elasticity control process, when being enforced, produces effects in different parts of the cloud service. These effects normally evolve in time and depend on workload characteristics, and on the actions within the elasticity control process enforced. Therefore, understanding the effects on the behavior of the cloud service is of utter importance for runtime decision-making process, when controlling cloud service elasticity. In this paper, we present a novel methodology and a framework for estimating and evaluating cloud service elasticity behaviors. To estimate the elasticity behavior, we collect information concerning service structure, deployment, service runtime, control processes, and cloud infrastructure. Based on this information, we utilize clustering techniques to identify cloud service elasticity behavior, in time, and for different parts of the service. Knowledge about such behavior is utilized within a cloud service elasticity controller to substantially improve the selection and execution of elasticity control processes. These elasticity behavior estimations are successfully being used by our elasticity controller, in order to improve runtime decision quality. We evaluate our framework with three real-world cloud services in different application domains. Experiments show that we are able to estimate the behavior in 89.5% of the cases. Moreover, we have observed improvements in our elasticity controller, which takes better control decisions, and does not exhibit control oscillations. © 2015 World Scientific Publishing Company. 24 3 Cited By :1

BACKGROUND: Streptococcus suis infection, an emerging zoonosis, is an increasing public health problem across South East Asia and the most common cause of acute bacterial meningitis in adults in Vietnam. Little is known of the risk factors underlying the disease. METHODS AND FINDINGS: A case-control study with appropriate hospital and matched community controls for each patient was conducted between May 2006 and June 2009. Potential risk factors were assessed using a standardized questionnaire and investigation of throat and rectal S. suis carriage in cases, controls and their pigs, using real-time PCR and culture of swab samples. We recruited 101 cases of S. suis meningitis, 303 hospital controls and 300 community controls. By multivariate analysis, risk factors identified for S. suis infection as compared to either control group included eating "high risk" dishes, including such dishes as undercooked pig blood and pig intestine (OR(1) = 2.22; 95%CI = [1.15-4.28] and OR(2) = 4.44; 95%CI = [2.15-9.15]), occupations related to pigs (OR(1) = 3.84; 95%CI = [1.32-11.11] and OR(2) = 5.52; 95%CI = [1.49-20.39]), and exposures to pigs or pork in the presence of skin injuries (OR(1) = 7.48; 95%CI = [1.97-28.44] and OR(2) = 15.96; 95%CI = [2.97-85.72]). S. suis specific DNA was detected in rectal and throat swabs of 6 patients and was cultured from 2 rectal samples, but was not detected in such samples of 1522 healthy individuals or patients without S. suis infection. CONCLUSIONS: This case control study, the largest prospective epidemiological assessment of this disease, has identified the most important risk factors associated with S. suis bacterial meningitis to be eating 'high risk' dishes popular in parts of Asia, occupational exposure to pigs and pig products, and preparation of pork in the presence of skin lesions. These risk factors can be addressed in public health campaigns aimed at preventing S. suis infection.

Currently, rich and diverse data types have been increasingly provided using the Data-as-a-Service (DaaS) model, a form of cloud computing services. However, data offered by DaaS are constrained by several data concerns that, if not automatically being reasoned properly, will lead to a wrong way of using them. In this paper, we support the assumption that data concerns should be explicitly modeled and specified in data contracts to support concern-aware data selection and utilization. Instead of relying on a specific definition of data contracts, we analyze contemporary data contracts and we present an abstract model for data contracts. Based on the abstract model, we propose several techniques for evaluating data contracts that can be integrated into data service selection and composition frameworks. We also illustrate our approach with some real world scenarios.

The increasing availability of massive information on the Web causes the need for information aggregation by filtering and ranking according to user's goals. In the last years both industrial and academic researchers have investigated the way in which quality of services can be described, matched, composed and monitored for service selection and composition. However, very few of them have considered the problem of evaluating and certifying the quality of the provided service information to reduce irrelevant information for service consumers, which is crucial to improve the efficiency and correctness of service composition and execution. This paper discusses several problems due to the lack of appropriate way to manage quality and context in service composition and execution, and proposes a research roadmap for reducing irrelevant service information based on context and quality aspects. We present a novel solution for dealing with irrelevant information about Web services by developing information quality metrics and by discussing experimental evaluations.

Traditional data quality engineering techniques, often used and deployed within a single enterprise environment, are inadequate to cope with the rapid change of data, with a multitude of quality degrees, to be used in contemporary business models. The emerging cloud computing paradigm could potentially offer high-quality, composable data and techniques, under the Software-as-a-Service (SaaS), Data-asa-Service (DaaS) and crowdsourcing models, for data quality engineering and data publishing. However, so far how to utilize the potential of cloud computing models for data quality engineering has not been discussed. In this paper, we analyze requirements of data quality engineering and quality-aware data publishing processes in the cloud and we provide a conceptual architecture utilizing and supporting the SaaS, DaaS and crowdsourcing models for the realization of such processes.

Currently, rich and diverse data types have been increasingly provided using the dataas- a-service (DaaS) model, a form of cloud computing services and the core element of data marketplaces. This facilitates the on-the-fly data composition and utilisation for several data-intensive applications in e-science and business domains. However, data offered by DaaS are constrained by several data concerns that, if not automatically being reasoned properly, will lead to a wrong way of using them. In this paper, we support the view that data concerns should be explicitly modelled and specified in data contracts to support concern-aware data selection and utilisation. We perform a detailed analysis of current techniques for data contracts in the cloud. Instead of relying on a specific representation of data contracts, we introduce an abstract model for data contracts that can be used to build different types of data contracts for specific types of data. Based on the abstract model, we propose several techniques for evaluating data contracts that can be integrated into data service selection and composition frameworks. We also illustrate our approach with some real-world scenarios and show how data contracts can be integrated into data agreement exchange services in the cloud.