Search

Abstract Online rumors are widespread and difficult to identify, which bring serious harm to society and individuals. To effectively detect and govern online rumors, it is necessary to conduct in-depth semantic analysis and understand the content features of rumors. This paper proposes a TFI domain ontology construction method, which aims to achieve semantic parsing and reasoning of the rumor text content. This paper starts from the term layer, the frame layer, and the instance layer, and based on the reuse of the top-level ontology, the extraction of core literature content features, and the discovery of new concepts in the real corpus, obtains the core classes (five parent classes and 88 subclasses) of the rumor domain ontology and defines their concept hierarchy. Object properties and data properties are designed to describe relationships between entities or their features, and the instance layer is created according to the real rumor datasets. OWL language is used to encode the ontology, Protégé is used to visualize it, and SWRL rules and pellet reasoner are used to mine and verify implicit knowledge of the ontology, and judge the category of rumor text. This paper constructs a rumor domain ontology with high consistency and reliability.

The integration density of silicon photonic integrated circuit (PIC) is ultimately constrained by various crosstalk mechanisms on the chip. Among them, the most prominent limiting factor is the thermal crosstalk due to the wide use of the thermo-optic effect. High-density silicon PICs strongly demand an advanced structure with better thermal crosstalk suppression ability than the traditional air isolation trench. Inspired by the thermal-metamaterial based on the scattering-cancellation method, we demonstrate a closed heat shield (CHS) structure on a silicon PIC chip, which can manipulate the thermal flux to bypass the temperature-sensitive silicon photonics components. The on-chip CHS structure is a bilayer cylindrical shell fabricated by the standard silicon photonics processing flow. Its outer and inner shell layers are formed by a 6-μm-wide interconnection metal and 4-μm-wide air trench, respectively. Plenty of temperature-sensitive micro-ring resonators inside the CHS are used to probe the temperature profile. The measurement results show that the CHS can reduce the local temperatures by 50%/44%/36% at the locations 29/41/83 μm away from the external heater. In contrast, the conventional air trench of the same dimension reduces the local temperatures by 32%/28%/21% at the same positions. In addition, the response time of the thermal field inside the CHS is around one-half of that in the conventional air trench. Furthermore, the simulation result indicates that if the outer shell of the CHS can contact with the silicon substrate by utilizing the through-silicon-via structure, the thermal crosstalk suppression ability can be improved significantly.

Glucocorticoid deficiency can lead to hypoglycemia, hypotension, and electrolyte disorders. Acute glucocorticoid deficiency under stress is very dangerous. Here, we present a case study of an elderly patient diagnosed with Sheehan's syndrome, manifesting secondary adrenal insufficiency and secondary hypothyroidism, managed with daily prednisone and levothyroxine therapy. She was admitted to our hospital due to acute non-ST segment elevation myocardial infarction. The patient developed nausea and limb twitching post-percutaneous coronary intervention, with subsequent diagnosis of hyponatremia. Despite initial intravenous sodium supplementation failed to rectify the condition, and consciousness disturbances ensued. However, administration of 50 mg hydrocortisone alongside 6.25 mg sodium chloride rapidly ameliorated symptoms and elevated blood sodium levels. Glucocorticoid deficiency emerged as the primary etiology of hyponatremia in this context, exacerbated by procedural stress during percutaneous coronary intervention. Contrast agent contributed to blood sodium dilution. Consequently, glucocorticoid supplementation emerges as imperative, emphasizing the necessity of stress-dose administration of glucocorticoid before the procedure. Consideration of shorter intervention durations and reduced contrast agent dosages may mitigate severe hyponatremia risks. Moreover, it is crucial for this patient to receive interdisciplinary endocrinologist management. In addition, Sheehan's syndrome may pose a risk for coronary atherosclerotic disease.

We mistakenly diagnosed a patient with pulmonary embolism as ST-elevation acute myocardial infarction. The case complained of chest pain with ST-elevation in leads AVR, III, AVF, V1, V3R, V4R and V5R. In addition, we also found ST- elevation AVR, III>ST- elevation AVF, QIII>QAVF and QV1. Multi-site pulmonary embolism can lead to type 2 myocardial infarction or special type of myocardial infarction with non-obstructive coronary arteries through multiple pathological mechanisms.

Ionic liquids (ILs) have been widely used in the pretreatment of biomass. However, the effects of residual ILs on the enzymolysis and fermentation of biomass are still unknown. Therefore, a large quantity of water-washing is usually followed after biomass pretreatment to eliminate the inhibition of residual ILs on subsequent hydrolysis and fermentation steps. In this work, the effect of choline glycine ([Ch][Gly]) concentration on the activity of cellulase and Bacillus sp. strain P38 was systematically investigated to explore the impacts of residual ILs on enzymolysis and fermentation. The results confirmed that the activities of them were almost not inhibited in low concentrations (less than 0.5 wt%) of [Ch][Gly]. Under optimal pretreatment conditions, the maximum cellulose digestibility was 99.23%. Enzymatic hydrolysate was suitable for L-lactic acid fermentation without appreciable inhibition, and the highest sugar-acid conversion rate of 96.33% was obtained by simplified detoxification. This work provides an economic route to produce fermentable sugar and L-lactic acid, which shows an industrial application prospect in lignocellulosic biorefinery.

We reviewed previously reported imaging features of coronary air embolism. We also reported an unusual image ‘rebound sign’ of proximal coronary artery air embolism for the first time, that is, with the regular contraction and relaxation of the heart, the contrast medium near the embolus rebounds regularly, and the smooth gas-liquid boundary is alternately clear and blurred. We conclude that myocardial bridge plays a role in the formation of ‘rebound sign.’ Abbreviations: CAE: Coronary Air Embolism; CAG: Coronary Angiography; RCA: Right Coronary Artery; LCA: Left Coronary Artery; LAD: Left Anterior Descending Coronary Artery; LCX: Left Circumflex Artery; MB: Myocardial Bridge; STEMI: ST Segment Elevation Myocardial Infarction; TOE: Transesophageal Echocardiography

We reported a patient with a fistula of the right coronary artery to the left ventricle, accompanied by dilation of the right coronary artery and persistent chest pain. This patient underwent surgical fistula closure surgery, but the fistula recurred. Persistent chest pain reappeared after encountering COVID-19 infection. We analyzed the mechanism of persistent myocardial ischemic chest pain caused by coronary artery fistula in this patient, the impact of surgery on the patient's disease, the possible mechanism of COVID-19 causing persistent ischemic chest pain in this patient, and the possible mechanism of metoprolol in alleviating myocardial ischemic chest pain in this patient.

Unmanned Aerial Vehicles (UAVs) have critical applications in various real-world scenarios, including mapping unknown environments, military reconnaissance, and post-disaster search and rescue. In these scenarios where communication infrastructure is missing, UAVs will form an ad hoc network and perform tasks in a distributed manner. To efficiently carry out tasks, each UAV must acquire and share global status information and data from neighbors. Meanwhile, UAVs frequently operate in extreme conditions, including storms, lightning, and mountainous areas, which significantly degrade the quality of wireless communication. Additionally, the mobility of UAVs leads to dynamic changes in network topology. Therefore, we propose a method that utilizes graph neural networks (GNN) to learn cooperative data dissemination. This method leverages the network topology relationship and enables UAVs to learn a decision policy based on local data structure, ensuring that all UAVs can recover global information. We train the policy using reinforcement learning that enhances the effectiveness of each transmission. After repeated simulations, the results validate the effectiveness and generalization of the proposed method.

BackgroundIn the early stage of COVID-19 epidemic, the Chinese mainland once effectively controlled the epidemic, but COVID-19 eventually spread faster and faster in the world. The purpose of this study is to clarify the differences in the epidemic data of COVID-19 in different areas and phases in Chinese mainland in 2020, and to analyze the possible factors affecting the occurrence and development of the epidemic.MethodsWe divided the Chinese mainland into areas I, I and III, and divided the epidemic process into phases I to IV: limited cases, accelerated increase, decelerated increase and containment phases. We also combined phases II and III as outbreak phase. The epidemic data included the duration of different phases, the numbers of confirmed cases, asymptomatic infections, and the proportion of imported cases from abroad.ResultsIn area I, II and III, only area I has a Phase I, and the Phase II and III of area I are longer. In Phase IV, there is a 17-day case clearing period in area I, while that in area II and III are 2 and 0 days, respectively. In phase III or the whole outbreak phase, the average daily increase of confirmed cases in area I was higher than that in areas II and III (P = 0.009 and P = 0.001 in phase III; P = 0.034 and P = 0.002 in the whole outbreak phase), and the average daily in-hospital cases were most in area I and least in area III (P = 0.000, P = 0.000, and P = 0.000 in phase III; P = 0.000, P = 0.000, and P = 0.009 in the whole outbreak phase). The average number of daily in-hospital COVID-19 cases in phase III was more than that in phase II in each area (P = 0.000, P = 0.000, and P = 0.001). In phase IV, from March 18, 2020 to January 1, 2021, the increase of confirmed cases in area III was higher than areas I and II (both P = 0.000), and the imported cases from abroad in Chinese mainland accounted for more than 55–61%. From June 16 to July 2, 2020, the number of new asymptomatic infections in area III was higher than that in area II (P = 0.000), while there was zero in area I. From July 3, 2020 to January 1, 2021, the increased COVID-19 cases in area III were 3534, while only 14 and 0, respectively, in areas I and II.ConclusionsThe worst epidemic areas in Chinese mainland before March 18, 2020 and after June 15, 2020 were area I and area III, respectively, and area III had become the main battlefield for Chinese mainland to fight against imported epidemic since March 18, 2020. In Wuhan, human COVID-19 infection might occur before December 8, 2019, while the outbreak might occur before January 16 or even 10, 2020. Insufficient understanding of COVID-19 hindered the implementation of early effective isolation measures, leading to COVID-19 outbreak in Wuhan, and strict isolation measures were effective in controlling the epidemic. The import of foreign COVID-19 cases has made it difficult to control the epidemic of area III. When humans are once again faced with potentially infectious new diseases, it is appropriate to first and foremost take strict quarantine measures as soon as possible, and mutual cooperation between regions should be explored to combat the epidemic.

Based on cascaded Mach-Zehnder interferometer (MZI) lattice filters, we demonstrate and compare silicon O-band 8-channel (de-)multiplexers with flat and Gaussian-like passbands for the 400 GBASE-LR8 norm. The 1×8 (de-)multiplexer with flat passbands exhibit insertion loss less than 2.5 dB, channel crosstalk lower than −11 dB, and 3 dB bandwidths of ∼3.9 nm. In contrast, the maximum insertion loss, the worst channel crosstalk, and 3 dB bandwidths of the 1×8 (de-)multiplexer with Gaussian-like passbands are 1.1 dB, −13 dB, and ∼2.9 nm, respectively. The comparison result indicates that the latter is more advisable choice. A theoretical calculation based on measured coupling ratios of stand-along directional couplers (DC) indicates that the performance deteriorations of practical filtering curves are mainly caused by dispersions and fabrication errors of the DCs.

Portrait matting is of vital importance for many applications such as portrait editing, background replacement, ecommerce demonstration, and augmented reality. The portrait matt can be accessed by predicting the α value of the original picture. Previous deep matting methods usually adopt a segmentation network to tackle portrait matting tasks. However, these traditional methods will introduce unpleasant blemishes in the matting results sometimes. The authors find that the key factor behind this phenomenon is how they model the matting problem. On the one hand, α value predicting can be modelled as a regression task. On the other hand, it can be viewed as a classification task of predicting background or foreground. To solve this problem, they explore different methods to model the nature of the α matting problem and propose a novel quantisation‐based adaption. Their method comes up with an α quantisation loss to achieve multi‐threshold filtering. Furthermore, they apply an α merging block to improve conventional regression methods. With their method, the gradient loss is reduced by 7.53% relatively, with mean square error and sum of absolute difference decreased by 14.7% relatively, leading to a more visually pleasant α matt in several segmentation backbones.

Fas knockout (Fas−/-) mice are a model for systemic lupus erythematosus (SLE) -like autoimmune syndromes. We aimed to induce atherosclerosis (AS) in Fas−/- mice. Sixteen male Fas−/- mice were included in the study, sex-matched C57B6/L (B6) and apolipoprotein E-knockout (ApoE−/-) mice were negative and positive AS controls, respectively. A silica collar was placed around the right carotid artery of each mouse to induce AS development. All mice were fed a 24-week high-fat diet, and investigated for AS lesions. We also compared the levels of blood lipid and glucose, serum tumor necrosis factor α (TNF-α), interleukin 6 (IL-6), anti-nuclear antibody (ANA) and anti-double-stranded deoxyribonucleic acid (anti-dsDNA) antibody in Fas−/- mice with those in B6 or ApoE−/- mice. All ApoE−/- and 6 Fas−/- but no B6 mice showed atherogenesis in right carotid artery. The carotid plaque contains more collagen and less lipid in Fas−/- than ApoE−/- mice. The levels of blood glucose, serum TNF-α, IL-6, ANA, and anti-dsDNA antibody were significantly higher in Fas−/- mice than those in B6 mice, the levels of serum TNF-α and blood glucose were significantly higher and the level of blood lipid was significantly lower in Fas−/- mice than those in ApoE−/- mice. Therefore, carotid AS can develop in Fas−/- mice. Fas−/- mice display higher levels of serum IL-6, TNF-α, ANA, and anti-dsDNA than B6 mice, higher levels of serum TNF-α and blood glucose and lower level of blood lipid than ApoE−/- mice, and less lipid and more collagen in AS plaque than ApoE−/- mice.

We demonstrate a thermally tunable four-channel add/drop demultiplexer based on silicon grating-assisted contra-directional couplers (contra-DCs) for 200 GHz dense wavelength division multiplexing (DWDM) systems. In order to obtain the narrow bandwidth and the rapid roll-off necessary for the DWDM, every wavelength channel is filtered dually by two identical contra-DCs cascaded in parallel. In addition, the phase apodization technique is utilized to suppress sidelobes. With these special designs, passbands of the four drop ports exhibit 3 and 20 dB bandwidths of 0.4 and 0.8 nm, respectively, while isolation between adjacent channels is better than 20 dB. Insertion losses at through and drop ports are 0.6 and ~2.6 dB, respectively. The maximal wavelength-tunable range is up to 25 nm.

Polarization light has been used as a key navigation cue by many animals under a stable celestial polarization pattern. Inspired by animals' unrivaled navigation skill, a bionic polarization sensor is designed previously to obtain the angle of polarization and degree of polarization. However, the accuracy and stability of the sensor are restricted by multi-source errors. In order to improve the accuracy of the sensor, in this paper, the multi-source errors are analyzed and modeled. A coarse calibration is presented to reduce the effect of inconsistency and interference. In addition, the extended Kalman filter and the unscented Kalman filter are designed to estimate the multi-source error or interference and compensate for the bionic polarization sensor. Finally, simulation and indoor experiments are conducted to validate the effectiveness of the proposed calibration method.

In this paper, a plasmonic Feynman gate based on suspended graphene nano-ribbon waveguides at THz wavelengths is proposed and numerically investigated. The proposed plasmonic Feynman gate is composed of two cascaded microring resonators. The resonant states of microring resonators are flexibly controlled by tuning chemical potentials of graphene in rings, and thus, the logic operation results of Feynman gate can be performed successfully. Compared to the plasmonic Feynman gate based on graphene nano-ribbons deposited on the Si substrate and the plasmonic Feynman gate based on graphene nano-ribbons deposited on the SiO2/Si substrate, our designed plasmonic Feynman gate can have better extinction ratio and crosstalk. Calculation results exhibit that the extinction ratio is larger than 15.39 dB and the crosstalk is lower than -15.55 dB when input logic states of our designed device are `00', `01', `10,' and `11'.

The Internet of Things (IoT) is the internetworking of terminal physical devices depends on application programming interfaces and sensors to be connected to the Internet for collecting and exchanging data. According to the characteristics of IoT, it can provide rich services for the terminal user. However, the centralized cloud computing-based storage architecture in IoT faces many challenges, such as data security, identity privacy, and high latency. To overcome these challenges, this paper introduces an IoT network storage architecture based on mobile edge computing, which not only achieves low-latency message response and computation offloading of the terminal user but also preserves identity-privacy of the terminal user. Afterward, we presented a novel non-interactive pairing-free ID-based proxy re-signature scheme (ID-PRS), which is a kernel technique to construct the aforementioned storage architecture. Compared with most of proxy re-signature schemes constructed based on traditional public-key-infrastructure, the proposed scheme avoids expensive pairing operation and intricate certificate-maintenance. And it is demonstrated to be provably secure under the classic security assumptions: integer factoring problem and the RSA assumption. Finally, in contrast to the other two ID-PRS schemes, the proposed ID-PRS scheme has more advantages in terms of security, functionability, and computation costs. Thus, it is very suitable to be applied to the resource-constrained mobile users.

A very challenging task for action recognition concerns how to effectively extract and utilize the temporal and spatial information of video (especially temporal information). To date, many researchers have proposed various spatial-temporal convolution structures. Despite their success, most models are limited in further performance especially on those datasets that are highly time-dependent due to their failure to identify the fusion relationship between the spatial and temporal features inside the convolution channel. In this paper, we proposed a lightweight and efficient spatial-temporal extractor, denoted as Channel-Wise Spatial-Temporal Aggregation block (CSTA block), which could be flexibly plugged in existing 2D CNNs (denoted by CSTANet). The CSTA Block utilizes two branches to model spatial-temporal information separately. In temporal branch, It is equipped with a Motion Attention Module (MA), which is used to enhance the motion regions in a given video. Then, we introduced a Spatial-Temporal Channel Attention (STCA) module, which could aggregate spatial-temporal features of each block channel-wisely in a self-adaptive and trainable way. The final experimental results demonstrate that the proposed CSTANet achieved the state-of-the-art results on EGTEA Gaze++ and Diving48 datasets, and obtained competitive results on Something-Something V1&V2 at the less computational cost.

The performance of voice-controlled systems is usually influenced by accented speech. To make these systems more robust, frontend accent recognition (AR) technologies have received increased attention in recent years. As accent is a high-level abstract feature that has a profound relationship with language knowledge, AR is more challenging than other language-agnostic audio classification tasks. In this paper, we use an auxiliary automatic speech recognition (ASR) task to extract language-related phonetic features. Furthermore, we propose a hybrid structure that incorporates the embeddings of both a fixed acoustic model and a trainable acoustic model, making the language-related acoustic feature more robust. We conduct several experiments on the AESRC dataset. The results demonstrate that our approach can obtain an 8.02% relative improvement compared with the Transformer baseline, showing the merits of the proposed method.

A recommendation system is often used to recommend items that may be of interest to users. One of the main challenges is that the scarcity of actual interaction data between users and items restricts the performance of recommendation systems. To solve this problem, multi-modal technologies have been used for expanding available information. However, the existing multi-modal recommendation algorithms all extract the feature of single modality and simply splice the features of different modalities to predict the recommendation results. This fusion method can not completely mine the relevance of multi-modal features and lose the relationship between different modalities, which affects the prediction results. In this paper, we propose a Cross-Modal-Based Fusion Recommendation Algorithm (CMBF) that can capture both the single-modal features and the cross-modal features. Our algorithm uses a novel cross-modal fusion method to fuse the multi-modal features completely and learn the cross information between different modalities. We evaluate our algorithm on two datasets, MovieLens and Amazon. Experiments show that our method has achieved the best performance compared to other recommendation algorithms. We also design ablation study to prove that our cross-modal fusion method improves the prediction results.

Face recognition/verification has received great attention in both theory and application for the past two decades. Deep learning has been considered as a very powerful tool for improving the performance of face recognition/verification recently. With large labeled training datasets, the features obtained from deep learning networks can achieve higher accuracy in comparison with shallow networks. However, many reported face recognition/verification approaches rely heavily on the large size and complete representative of the training set, and most of them tend to suffer serious performance drop or even fail to work if fewer training samples per person are available. Hence, the small number of training samples may cause the deep features to vary greatly. We aim to solve this critical problem in this paper. Inspired by recent research in scene domain transfer, for a given face image, a new series of possible scenarios about this face can be deduced from the scene semantics extracted from other face individuals in a face dataset. We believe that the “scene” or background in an image, that is, samples with more different scenes for a given person, may determine the intrinsic features among the faces of the same individual. In order to validate this belief, we propose a Bayesian scene-prior-based deep learning model in this paper with the aim to extract important features from background scenes. By learning a scene model on the basis of a labeled face dataset via the Bayesian idea, the proposed method transforms a face image into new face images by referring to the given face with the learnt scene dictionary. Because the new derived faces may have similar scenes to the input face, the face-verification performance can be improved without having background variance, while the number of training samples is significantly reduced. Experiments conducted on the Labeled Faces in the Wild (LFW) dataset view #2 subset illustrated that this model can increase the verification accuracy to 99.2% by means of scenes’ transfer learning (99.12% in literature with an unsupervised protocol). Meanwhile, our model can achieve 94.3% accuracy for the YouTube Faces database (DB) (93.2% in literature with an unsupervised protocol).