Search

Background: The contribution of sex as an independent risk factor for cardiovascular disease still remains controversial. The present study investigated the impact of sex on long-term clinical outcomes in Korean acute myocardial infarction (AMI) patients with a history of current smoking on admission after drug-eluting stents (DESs). Methods: A total of 12,565 AMI patients (male: n = 11767 vs. female: n = 798) were enrolled. Major adverse cardiac events (MACEs) comprising all-cause death, recurrent myocardial infarction (Re-MI), and any repeat revascularization were the primary outcomes that were compared between the two groups. Probable or definite stent thrombosis (ST) was the secondary outcome. Results: After adjustment, the early (30 days) cumulative incidences of MACEs (adjusted hazard ratio [aHR]: 1.457; 95% confidence interval [CI]: 1.021–2.216; p = 0.035) and all-cause death (aHR: 1.699; 95% CI: 1.074–2.687; p = 0.023) were significantly higher in the female group than in the male group. At 2 years, the cumulative incidences of all-cause death (aHR: 1.561; 95% CI: 1.103–2.210; p = 0.012) and Re-MI (aHR: 1.880; 95% CI: 1.089–2.974; p = 0.022) were significantly higher in the female group than in the male group. However, the cumulative incidences of ST were similar between the two groups (aHR: 1.207; 95% CI: 0.583–2.497; p = 0.613). Conclusions: The female group showed worse short-term and long-term clinical outcomes compared with the male group comprised of Korean AMI patients with a history of current smoking after successful DES implantation. However, further studies are required to confirm these results.

Scene parsing, or semantic segmentation, aims at labeling all pixels in an image with the predefined categories of things and stuff. Learning a robust representation for each pixel is crucial for this task. Existing state-of-the-art (SOTA) algorithms employ deep neural networks to learn (discover) the representations needed for parsing from raw data. Nevertheless, these networks discover desired features or representations only from the given image (content), ignoring more generic knowledge contained in the dataset. To overcome this deficiency, we make the first attempt to explore the meaningful supportive knowledge, including general visual concepts (i.e., the generic representations for objects and stuff) and their relations from the whole dataset to enhance the underlying representations of a specific scene for better scene parsing. Specifically, we propose a novel supportive knowledge mining module (SKMM) and a knowledge augmentation operator (KAO), which can be easily plugged into modern scene parsing networks. By taking image-specific content and dataset-level supportive knowledge into full consideration, the resulting model, called knowledge augmented neural network (KANN), can better understand the given scene and provide greater representational power. Experiments are conducted on three challenging scene parsing and semantic segmentation datasets: Cityscapes, Pascal-Context, and ADE20K. The results show that our KANN is effective and achieves better results than all existing SOTA methods.

This article conducts a comparative analysis to investigate the effects of different classification algorithms and structural proximity indexes on the performance of the supervised link prediction approach to anticipating technology convergence at different forecast horizons. For this, we identify relationships between technologies of interest for different time periods and compute 10 structural proximity indexes among unconnected technologies at each period. We develop a set of classification models that identify potential convergence among unconnected technologies where each model is configured differently by a classification algorithm and a combination of the proximity indexes. We compare the performance of the classification models to investigate effective combinations of classification algorithms and proximity indexes at different forecast horizons. The empirical analysis on Wikipedia articles about artificial intelligence technology indicates that random forest outperforms others in short-term forecasting while support vector machine outperforms others in mid-term forecasting. We also identify structural proximity indexes that produce higher performance when combined with the most effective algorithm at each forecast horizon. The results of this article are expected to offer guidelines for choosing classification algorithms and indexes when applying the supervised link prediction approach in anticipating technology convergence.

Motivated by the practices of technology investment in the new energy vehicle manufacturing industry, this article investigates an outsourcing supply chain in which an original brand manufacturer (OBM) possessing the core new energy technology R&D outsources production to a competitive manufacturer (CM). To cater to market demands and increase market competitiveness, the CM may actively invest in technology with the OBM or by himself, or not, i.e., Strategies JI, OI, and NI, respectively. In particular, owing to technology spillover effect, the CM could use the technology inferior to the OBM's to produce his own-branded products in Strategy NI. Our results reveal that the OBM consistently profits the most in Strategy JI. However, the CM's preference strategy is influenced by the competition intensity, technology spillover level, consumers’ product preference, and cost-sharing ratio. In specific, when both the competition intensity and technology spillover are weak or the competition intensity is strong for arbitrary technology spillover level, the CM may prefer Strategy JI or OI depending on consumers’ product preference and cost-sharing ratio and never prefers Strategy NI. However, when the competition intensity is weak but the technology spillover level is strong, the CM may prefer Strategy NI in certain situations. Furthermore, our study extends the model to involve variable production costs, and reveals that the main results derived in the basic model still hold.

Objectives This study was to explore the potential relationship between the fibrinogen-to-albumin ratio (FAR) and the presence and severity of coronary artery disease (CAD) in stage 3–5 predialysis chronic kidney disease (CKD) patients. Design This study included 978 patients undergoing coronary angiography (CAG). CAD was defined as the presence of obstructive stenosis > 50% of the lumen diameter in any of the four main coronary arteries. Gensini scores (GSs), left main coronary artery (LMCA) and three-vessel coronary artery disease (TVD) were used to elevate the severity of CAD. Results The adjusted odds ratios of CAD were 3.059 (95% CI: 1.859–5.032) and 2.670 (95% CI: 1.605–4.441) in the third and fourth quartiles of FAR compared with the first quartile, respectively. Among 759 patients diagnosed with CAD, multivariate logistic regression analysis showed that FAR (at the 0.01 level) was significantly positively associated with the presence of LMCA (adjusted OR = 1.177, 95% CI 1.067–1.299, P = 0.001) or TVD (adjusted OR = 1.154, 95% CI 1.076–1.238, P Conclusions FAR levels were independently associated with the presence and severity of CAD in stage 3–5 predialysis CKD patients.

The cell-type continuous electromagnetic radiation system is a demonstration device capable of generating high-power millimeter electromagnetic waves of a specific wavelength and observing their effects on living organisms. It irradiates a biological sample placed in a 30 × 30 × 50 cm3 cell with electromagnetic waves in the 3.15-mm-wavelength region (with an output of ≥1 W) and analyzes the temperature change of the sample. A vacuum electronic device-based coupled-cavity backward-wave oscillator converts the electron energy of the electron beam into radiofrequency (RF) energy and radiates it to the target through an antenna, increasing the temperature through the absorption of RF energy in the skin. The system causes pain and ultimately reduces combat power. A cell-type continuous electromagnetic radiation system consisting of four parts—an electromagnetic-wave generator, a high-voltage power supply, a test cell, and a system controller—generates an RF signal of ≥1 W in a continuous waveform at a 95-GHz center frequency, as well as a chemical solution with a dielectric constant similar to that of the skin of a living organism. An increase of 5 °C lasting approximately 10 s was confirmed through an experiment.

Objective Genetically engineered pigs are thought to be an alternative organ source for patients in end-stage heart failure unable to receive a timely allograft. However, cardiac xenografts exhibit growth and diastolic heart failure within 1 month after transplantation. Grafts function for up to 6 months, but only after administration of temsirolimus and afterload-reducing agents to reduce this growth. In this study we investigated the growth and hemodynamics of growth hormone receptor (GHR) knockout xenografts, without the use of adjuncts to prevent intrinsic graft growth after transplantation. Methods Genetically engineered pig hearts were transplanted orthotopically into weight-matched baboons between 15 and 30 kg, using continuous perfusion preservation before implantation (n = 5). Xenografts included knockout of carbohydrate antigens and knockin of human transgenes for thromboregulation, complement regulation, and inflammation reduction (grafts with intact growth hormone, n = 2). Three grafts contained the additional knockout of GHR (GHR knockout grafts; n = 3). Transthoracic echocardiograms were obtained twice monthly and comprehensively analyzed by a blinded cardiologist. Hemodynamics were measured longitudinally after transplantation. Results All xenografts demonstrated life-supporting function after transplantation. There was no difference in intrinsic growth, measured using septal and posterior wall thickness and left ventricular mass, on transthoracic echocardiogram out to 1 month in either GHR knockout or GHR intact grafts. However, hypertrophy of the septal and posterior wall was markedly elevated by 2 months post transplantation. There was minimal hypertrophy out to 6 months in GHR knockout grafts. Physiologic mismatch was present in all grafts after transplantation, which is largely independent of growth. Conclusions Xenografts with GHR knockout show reduced post-transplantation xenograft growth using echocardiography >6 months after transplantation, without the need for other adjuncts.

OBJECTIVE To clarify whether systemic lymph node dissection (LND) influences the safety of surgery and the survival of patients with locally advanced esophageal squamous cell carcinoma (ESCC) after neoadjuvant chemoradiotherapy (nCRT). SUMMARY BACKGROUND DATA Prognostic impact of systemic lymphadenectomy during surgery after nCRT for ESCC is still uncertain and requires clarification. METHODS This is a secondary analysis of NEOCRTEC5010 trial which compared nCRT followed by surgery versus surgery alone for locally advanced ESCC. Relationship between number of LND and perioperative, recurrence and survival outcomes were analyzed in the nCRT group. RESULTS Three-year overall survival was significantly better in the nCRT group than the S group (75.2% vs 61.5%; P=0.011). In the nCRT group, greater number of LND was associated with significantly better overall survival (HR, 0.358; P < 0.001) and disease-free survival (HR, 0.415; P=0.001), but without any negative impact on postoperative complications. Less LND (< 20 vs ≥ 20) was significantly associated with increased local recurrence (18.8% vs 5.2%, P=0.004) and total recurrence rates (41.2% vs 25.8%, P=0.027). Compared to patients with persistent nodal disease, significantly better survival was seen in patients with complete response and with LND ≥ 20, but not in those with LND < 20. CONCLUSIONS Systemic lymph node dissection does not increase surgical risks after nCRT in ESCC patients. And it is associated with better survival and local disease control. Therefore, systemic lymphadenectomy should still be considered as an integrated part of surgery after nCRT for ESCC.

Background: Chorea-acanthocytosis (ChAc) is the most common type of neuroacanthocytosis syndromes. Characteristic movement disorders of ChAc are choreiform movements affecting both trunk and extremities. Acanthocytosis in peripheral blood smear, elevated serum creatine kinase, atrophy of heads of caudate nuclei and dilation of the anterior horn of the lateral ventricles in magnetic resonance imaging could assist the diagnosis of ChAc. Objective: We aimed to report on the use of acupuncture to successfully improve ChAc symptoms. Method: A patient with definite ChAc was admitted, who had suffered from involuntary tongue protrusion for about 10 years. Acupuncture treatment was administrated for 3 times a week for 2 months. The chorea tremor control area, Baihui (GV20), Sishencong (EX-HN1), Shenting (GV24), Benshen (GB13, bilateral), Yintang (GV29), Neiguan (PC6, bilateral), Tongli (HT5, bilateral), Zusanli (ST36, bilateral), Sanyinjiao (SP6, bilateral), Dicang (ST4, bilateral), Chengjiang (CV24), Lianquan (CV23), Jinjin (EX-HN12) and Yuye (EX-HN13) were selected as acupunture points. Results: Previous drug dosage was reduced and the frequency of involuntary tongue protrusion was significantly reduced. Other clinical symptoms were also well controlled. Peripheral blood smear still indicated an increased proportion of red lineage, but blood analyses revealed improvement at follow-up. Conclusions: For patients who do not response well to conventional medical treatments, acupuncture might be used as an alternative treatment for symptoms related to ChAc.

The satisfaction of a software requirement is commonly stated as a Boolean value, that is, a security requirement is either satisfied (true) or not (false). However, a discrete Boolean value to measure the satisfaction level of a security requirement by deployed mechanisms is not very useful. Rather, it would be more effective if we could quantify the level of satisfaction of security requirements on a continuous scale. We propose an approach to achieve this for cloud software systems based on relationships between defense strength, exploitability of vulnerabilities, and attack severity. We extend the concept of entailment relationship from the field of requirements engineering with the satisfiability aspects of security requirements. The proposed approach enables us to systematically structure security concepts into three sets of related descriptions to quantify the satisfaction level of security requirements with the deployed security solutions. To demonstrate the feasibility of the proposed approach, we evaluate the approach in a case study. As a result, security administrators are able to deploy more effective and appropriate security solutions based on their assessment.

Cloud service providers typically provide different types of virtual machines (VMs) to cloud users with various requirements. Thanks to its effectiveness and fairness, auction has been widely applied in this heterogeneous resource allocation. Recently, several strategy-proof combinatorial cloud auction mechanisms have been proposed. However, they fail to protect the bid privacy of users from being inferred from the auction results. In this paper, we design a differentially private combinatorial cloud auction mechanism (DPCA) to address this privacy issue. Technically, we employ the exponential mechanism to compute a clearing unit price vector with a probability proportional to the corresponding revenue. We further improve the mechanism to reduce the running time while maintaining high revenues, by computing a single clearing unit price, or a subgroup of clearing unit prices at a time, resulting in the improved mechanisms DPCA-S and its generalized version DPCA-M, respectively. We theoretically prove that our mechanisms can guarantee differential privacy, approximate truthfulness and high revenue. Extensive experimental results demonstrate that DPCA can generate near-optimal revenues at the price of relatively high time complexity, while the improved mechanisms achieve a tunable trade-off between auction revenue and running time., 12 pages, 6 figures

The scarcity of labels is common and great challenge in traditional supervised learning. Semi-supervised learning (SSL) leverages unlabeled samples to alleviate the absence of label information. Similar with annotation, label proportion is another type of prior information and plays a significant role in classification tasks. Compared with the acquisition of labels, label proportion can be obtained more easily. For example, only a small number of patients have been diagnosed with or not with cancers in hospital database, while the proportion with cancer can be generally estimated by historical records. How to incorporate such prior information of label proportion is crucial but rarely studied in literature. Traditional SSL methods often ignore this prior information and will lead to performance degradation inevitably. To solve this problem, we propose a novel SSL with Label Proportion (SSLLP). Our approach encourages to preserve label consistency and label proportion by imposing the cardinality bound constraints. Our formulated problem equals to a mixed-integer constrained submodular minimization and it is difficult to be solved directly. Therefore, we transformed the original problem into a convex one by Lov $\acute{\text{a}}$ sz extension and designed an efficient solving algorithm. Extensive experimental results present the improved performance of our method over several state-of-the-art methods.

The world is entering a new era with ubiquitous connectivity among billions of humans and machines, i.e., the 6G massive Internet of Things (IoT). Radar and communication need to be integrated into this system to achieve target detection and massive connectivity. In this work, we first develop a shape-adaptive antenna array composed of multiple sub-arrays spaced at regular intervals. The shape-adaptive antenna array uses flexible material that enables changes in the physical structure to achieve adaptive gain. Subsequently, we investigate the micro-movement of IoT targets, such as the micro-motion characteristics and micro-Doppler effects of rotary-wing drones. The simulation results demonstrate the effectiveness and efficiency of the proposed schemes for detecting UAVs in an IoT scenario.

Thyroid cancer is not among the top cancers in terms of diagnosis or mortality but it still ranks fifth among the cancers diagnosed in women. Infact, women are more likely to be diagnosed with thyroid cancer than the males. The burden of thyroid cancer has dramatically increased in last two decades in China and, in the United States, it is the most diagnosed cancer in young adults under the age of twenty-nine. All these factors make it worthwhile to fully understand the pathogenesis of thyroid cancer. Towards this end, microRNAs (miRNAs) have constantly emerged as the non-coding RNAs of interest in various thyroid cancer subtypes on which there have been numerous investigations over the last decade and half. This comprehensive review takes a look at the current knowledge on the topic with cataloging of miRNAs known so far, particularly related to their utility as epigenetic signatures of thyroid cancer progression and metastasis. Such information could be of immense use for the eventual development of miRNAs as therapeutic targets or even therapeutic agents for thyroid cancer therapy.

Region-based methods are currently achieving state-of-the-art performance for monocular 3D object tracking. However, they are still prone to fail in cases of partial occlusions and ambiguous colors. We propose a novel region-based method to tackle these problems. The key idea is to derive a pixel-wise weighted region-based cost function using contour constraints. Firstly, we propose a novel region-based cost function using search lines around the object contour, which is more efficient than previous region-based cost functions using signed distance transform, and in the meantime can deal with partial occlusions and ambiguous colors more effectively. Secondly, we propose an optimal searching strategy to search the object contour points in cluttered scenes, and then use the object contour points to detect partial occlusions and ambiguous colors. Thirdly, we propose a pixel-wise weight function based on color and distance constraints of the object contour points, and integrate it into the proposed region-based cost function to reduce the negative impact of partial occlusions and ambiguous colors. We verify the effectiveness and efficiency of our method on challenging public datasets. Experiments demonstrate that our method outperforms the recent state-of-the-art region-based methods in complex scenarios, especially in the presence of partial occlusions and ambiguous colors.

Liver cancer as the main leading cancer has caused heavy burdens globally. The prognosis of liver cancer is closely related with postoperative nutrition support. Corn oligopeptides (COPs) are protein hydrolysates produced by enzymatic treatments, which have shown potential bioactivities, such as inhibiting angiotensin I-converting enzyme, resisting lipid peroxidation and anti-oxidant. However, the correlation between COPs and liver cancer patients is still unknown and the potential mechanism of COPs on liver cancer is unclear as well. The aim of this study was to assess effects of 7-day intervention of COPs after surgery on liver function and serum metabolic profiles of liver cancer patients.Patients were assigned into COPs intervention group (n=50) and control group (n=91) for 7 days. Investigations were scheduled at 1Seven-day supplementation of COPs on early post-surgery liver cancer patients down-regulated levels of alanine aminotransferase, aspartate aminotransferase, total bilirubin, direct bilirubin and up-regulated prothrombin time activity and prealbumin levels. LC/MS analysis revealed metabolic signatures including regulation of 16 metabolites, which was closely related with two metabolic pathways (nicotinate and nicotinamide metabolism, fatty acid metabolism).COPs supplementation has displayed the potentials on alleviating the injury of liver function and it may be due to regulation of fatty acid metabolism, nicotinate and nicotinamide metabolism, lipid peroxidation and anti-inflammatory action. More researches are warranted in future to confirm the exact mechanisms.

Intelligent reflecting surface (IRS) has received widespread attention because it can significantly improve the performance of wireless networks. The X-duplex is a more flexible duplex mode, which can decrease a small sum rate in exchange for a higher quality-of-service performance. In this article, an IRS transmission scheme based on X-duplex is investigated to enhance the flexibility and freedom of communication design. We present the X-duplex IRS with the simultaneous wireless information and power transfer, and establish a power consumption model based on real application scenarios. An alternating optimization algorithm is proposed to maximize the energy efficiency (EE) of a multi-input single-output (MIMO) network assisted by the X-duplex IRS. As for the X-duplex relay, we explore an algorithm based on the golden section to optimize the EE. Simulation results comprehensively evaluate the performance of the two auxiliary transmission methods. The X-duplex IRS has higher EE and weighted sum rate (WSR) than the X-duplex relay when sufficient transmission power is available. However, the IRS cannot entirely replace the traditional relay. The X-duplex relay with self-interference energy harvesting achieves a higher WSR and EE when the access point energy is limited or the transmission distance is short.

This study aims to find out the possible optimal therapy and assess the prognosis properly for patient with spontaneous rupture of hepatocellular carcinoma (HCC).Propensity score matching (PSM) analysis was used to study the data from 325 patients with ruptured HCC (RHCC) and 2,291 patients with non-RHCC.The incidence and hospital mortality of RHCC were 5.1% and 0.8% respectively, with a median overall survival (OS) time of 17 months. There was no difference between ruptured and non-RHCC patients undergoing conservation treatment in terms of OS. Trans-arterial embolization (TAE) was carried out in 69 (21.2%) cases with RHCC, with a median OS of 7 months, which was no difference from that of non-RHCC (pre- and post-PSM). One hundred and sixty-nine (52.0%) RHCC cases underwent one-stage hepatectomy, with a median OS and disease-free survival (DFS) of 30 and 6 months respectively, which were shorter than that of non-RHCC (post-PSM). TAE plus two-stage hepatectomy was performed in 30 RHCC cases, with a median OS and DFS of 28 and 10 months respectively; these outcomes were better than that from RHCC patients undergoing TAE alone or one-stage hepatectomy (post-PSM), which were no difference from that of non-RHCC patients undergoing hepatectomy. The risk of death for RHCC patient undergoing one-stage hepatectomy is 1.545 times higher than that of one undergoing TAE + two-stage hepatectomy.TAE plus two-stage hepatectomy might be the optimal treatment for RHCC patient. Under the premise of the same pathological properties, there is no difference in prognosis between ruptured and non-RHCC patients if the therapy is appropriate.

Introduction Legume crops are an important source of protein and oil for human health and in fixing atmospheric N2 for soil enrichment. With an objective to accelerate much-needed genetic analyses and breeding applications, draft genome assemblies were generated in several legume crops; many of them are not high quality because they are mainly based on short reads. However, the superior quality of genome assembly is crucial for a detailed understanding of genomic architecture, genome evolution, and crop improvement. Objectives Present study was undertaken with an objective of developing improved chromosome-length genome assemblies in six different legumes followed by their systematic investigation to unravel different aspects of genome organization and legume evolution. Methods We employed in situ Hi-C data to improve the existing draft genomes and performed different evolutionary and comparative analyses using improved genome assemblies. Results We have developed chromosome-length genome assemblies in chickpea, pigeonpea, soybean, subterranean clover, and two wild progenitor species of cultivated groundnut (A. duranensis and A. ipaensis). A comprehensive comparative analysis of these genome assemblies offered improved insights into various evolutionary events that shaped the present-day legume species. We highlighted the expansion of gene families contributing to unique traits such as nodulation in legumes, gravitropism in groundnut, and oil biosynthesis in oilseed legume crops such as groundnut and soybean. As examples, we have demonstrated the utility of improved genome assemblies for enhancing the resolution of “QTL-hotspot” identification for drought tolerance in chickpea and marker-trait associations for agronomic traits in pigeonpea through genome-wide association study. Genomic resources developed in this study are publicly available through an online repository, ‘Legumepedia’. Conclusion This study reports chromosome-length genome assemblies of six legume species and demonstrates the utility of these assemblies in crop improvement. The genomic resources developed here will have significant role in accelerating genetic improvement applications of legume crops.

The state-of-the-art optical see-through head-mounted displays (OST-HMD) for augmented reality applications lack the ability to render correct light interaction behavior between digital and physical objects, known as mutual occlusion capability. This paper presents a novel optical architecture for enabling a compact, high performance, occlusion-capable optical see-through head-mounted display (OCOST-HMD) with correct, pupil-matched viewing perspective. The proposed design utilizes a single-layer, double-pass architecture, offering a compact OCOST-HMD solution that is capable of rendering per-pixel mutual occlusion, a correctly pupil-matched viewing between virtual and real views, and a very wide see-through field of view (FOV). Based on this architecture, we demonstrate a design embodiment and a compact prototype implementation. The prototype offers a virtual display with an FOV of 34 by 22, an angular resolution of 1.06 arc minutes per pixel, and an average image contrast greater than 40% at the Nyquist frequency of 53 cycles/mm. Further, the prototype system affords a wide see-though FOV of 90 by 50, within which about 40 diagonally is occlusion-enabled, along with an angular resolution of 1.0 arc minutes, comparable to a 20/20 vision and a dynamic range greater than 100:1. Lastly, we composed a quantitative color study that compares the effects of occlusion between a conventional HMD system and our OCOST-HMD system and the resulting response exhibited in different studies.

Broadcasting is an essential operation for the source node to disseminate the message to all other nodes in the network. Unfortunately, the problem of Minimum Latency Broadcast Scheduling (MLBS) in duty-cycled wireless networks is not well studied. In existing works, the construction of broadcast tree and the scheduling of transmissions are conducted separately, which may result in a bad-structured broadcast tree and then a large latency is obtained even using the optimal scheduling method. Thus, the MLBS problem in duty-cycled wireless networks without above limitation is investigated in this paper. Firstly, a two step scheduling algorithm is proposed to construct the broadcast tree and compute a collision-free schedule simultaneously. This is the first work that can integrate these two kinds of operations together. Secondly, a novel transmission mode, i.e., concurrent broadcasting, is first introduced for wireless networks and several techniques are designed to further improve the broadcast latency. Thirdly, through designing some special data structures, the continuous and all-to-all broadcasting algorithms are also proposed, in which all the nodes do not need to wait for any period of time to begin a new transmission. Finally, the theoretical analysis and experimental results demonstrate the efficiency of the proposed algorithms.

The channel state information (CSI) is essential for the base station (BS) to schedule user equipments (UEs) and efficiently manage the radio resources. Hence, the BS requests UEs to regularly feed back the CSI. However, frequent CSI reporting causes large signaling overhead. To reduce the feedback overhead, we propose two machine learning-based approaches to adjust the CSI feedback interval. We use a deep neural network and reinforcement learning (RL) to decide whether an UE feeds back the CSI. Simulation results show that the RL-based approach achieves the lowest mean squared error while reducing the number of CSI feedback transmissions.

Complete tracheal obliteration after tracheostomy remains a considerable challenge for otolaryngologists and pulmonologists. Here, we report for the first time a novel method of interventional bronchoscopy to successfully recanalize complete tracheal obliteration. Three patients with suprastomal tracheal obliteration and tracheostomy dependence were referred to our center for further management. Using interventional bronchoscopy, a TBNA needle was retrogradely inserted from the stoma to locate the original passage through the occlusion, and then its stylet was left as a guide wire for the sequential dilations. Once the tracheal lumen was restored, endoprosthesis would be implanted to maintain the airway patency. All cases achieved successful recanalization with effortless breathing after the treatment and restored phonation. Bronchoscopic retrograde recanalization using a TBNA needle is a promising and effective treatment for complete tracheal obliteration.

How to generate more revenues is crucial to cloud providers. Evidences from the Amazon cloud system indicate that ``dynamic pricing'' would be more profitable than ``static pricing.'' The challenges are: How to set the price in real-time so to maximize revenues? How to estimate the price dependent demand so to optimize the pricing decision? We first design a discrete-time based dynamic pricing scheme and formulate a Markov decision process to characterize the evolving dynamics of the price-dependent demand. We formulate a revenue maximization framework to determine the optimal price and theoretically characterize the ``structure'' of the optimal revenue and optimal price. We apply the Q-learning to infer the optimal price from historical transaction data and derive sufficient conditions on the model to guarantee its convergence to the optimal price, but it converges slowly. To speed up the convergence, we incorporate the structure of the optimal revenue that we obtained earlier, leading to the VpQ-learning (Q-learning with value projection) algorithm. We derive sufficient conditions, under which the VpQ-learning algorithm converges to the optimal policy. Experiments on a real-world dataset show that the VpQ-learning algorithm outperforms a variety of baselines, i.e., improves the revenue by as high as 50% over the Q-learning, speedy Q-learning, and adaptive real-time dynamic programming (ARTDP), and by as high as 20% over the fixed pricing scheme.

Controversial results have been reported about the association of calcium, magnesium, and phosphorus and stroke risk, but none in China. To investigate the association between dietary calcium, magnesium, phosphorus, and stroke incidence in Chinese adults, we collected data from the China Health and Nutrition Survey (CHNS) from 2004 to 2011, including 6411 participants aged 45–79 years and free of stroke at baseline. Diet was assessed by interviews combining 3-d 24-h food recalls and household food inventory weighing at each survey round. The stroke incident was identified based on the validated self-report. Multivariate Cox regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). For 32,024 person-years of follow-up, 179 stroke cases were documented. After adjustment for major lifestyle and dietary risk factors, calcium intake was positively associated with reduced stroke risk, and the HR of stroke comparing extreme quartiles was 0.53 (95% CI 0.29–0.96, Ptrend = 0.03). In further stratified analyses, significant heterogeneity across sex strata was found (Pinteraction = 0.03). Dietary calcium intake among men was more inversely related to stroke, with HRs being 0.33 (95% CI 0.15–0.76, Ptrend = 0.02), compared to 1.24 (95% CI 0.46–3.35, Ptrend = 0.89) among women. However, no significant association between stroke and magnesium or phosphorus was revealed. Our findings suggest that higher dietary calcium intake was associated with a lower risk of stroke in Chinese adults, particularly in men.

Purpose Portal hypertension due to cirrhosis is common among patients with hepatocellular carcinoma (HCC). This study aimed to compare the outcomes of partial hepatectomy in patients with HCC and clinically significant portal hypertension (CSPH) with or without concurrent splenectomy and esophagogastric devascularization (CSED). Patients and methods From a multicenter database, patients with HCC and CSPH who underwent curative-intent hepatectomy were identified. Postoperative morbidity and mortality, and long-term overall survival (OS) were compared in patients with and without CSED before and after propensity score matching (PSM). Results Of the 358 enrolled patients, 86 patients underwent CSED. Before PSM, the postoperative 30-day morbidity and mortality rates were comparable between the CSED and non-CSED group (both P > 0.05). Using PSM, 81 pairs of patients were created. In the PSM cohort, the 5-year OS rate of the CSED group were significantly better than the non-CSED group (52.9%vs.36.5%, P=0.046). The former group had a significantly lower rate of variceal bleeding on follow-up (7.4%vs.21.7%, P=0.014). On multivariate analysis, CSED was associated with significantly better OS (HR: 0.39, P Conclusion Hepatectomy and CSED can safely be performed in selected patients with HCC and CSPH, which could improve postoperative prognosis by preventing variceal bleeding, and prolonging long-term survival.

To report three-decade changes of clinical characteristics, progress of treatments, and risk factors associated with mortality and enucleation in patients with retinoblastoma in China.Retrospective cohort study.This multicenter study included 2552 patients diagnosed with retinoblastoma in 38 medical centers in 31 provinces in China from 1989 to 2017, with follow-up data. Kendall's tau-b value was used to describe correlation coefficients between the three eras (between 1989 and 2008, between 2009 and 2013, and between 2014 and 2017) and clinical or demographic features. Hazard ratios and odds ratios were applied to measure risk factors.A total of 324 (13%) patients died and 1414 (42%) eyes were removed. The 1-year, 3-year, and 5-year overall survival rates were 95%, 86%, and 83%, respectively. Patients were diagnosed at a better stage by International Classification for Retinoblastoma over time (Kendall's tau-b value = -0.084, P.001). Pathological risk factors were also observed less in recent eras. New conservative therapies were adopted and used in more patients. The eye removal rate gradually decreased (Kendall's tau-b value = -0.167, P.001). The overall survival rates were 81%, 83%, and 91% in the three eras. By multivariate Cox regression, bilateral tumors and extraocular extension were identified as risk factors for death. Among intraocular disease, Group E indicated higher risk of mortality. By multivariate logistics regression, unilateral tumors, earlier era of diagnosis, and extraocular extension were risk factors for eye salvage failure. Among intraocular retinoblastoma, Groups D and E had higher risk of eye salvage failure.Patients were diagnosed at an earlier stage in recent eras. Conservative therapies, including intra-arterial chemotherapy, were increasingly being used. The above changes may contribute to the decreasing enucleation rate. Although no significant impact was identified on the mortality by the three eras, a decreasing trend was shown.

Recently, functional homotopy (FH) architecture, defined as robust functional connectivity (FC) between homotopic regions, has been frequently reported to be altered in MDD patients (MDDs) but with divergent locations.In this study, we obtained resting-state functional magnetic resonance imaging (R-fMRI) data from 1004 MDDs (mean age, 33.88 years; age range, 18-60 years) and 898 matched healthy controls (HCs) from an aggregated dataset from 20 centers in China. We focused on interhemispheric function integration in MDDs and its correlation with clinical characteristics using voxel-mirrored homotopic connectivity (VMHC) devised to inquire about FH patterns.As compared with HCs, MDDs showed decreased VMHC in visual, motor, somatosensory, limbic, angular gyrus, and cerebellum, particularly in posterior cingulate gyrus/precuneus (PCC/PCu) (false discovery rate [FDR] q 0.002, z = -7.07). Further analysis observed that the reduction in SMG and insula was more prominent with age, of which SMG reflected such age-related change in males instead of females. Besides, the reduction in MTG was found to be a male-special abnormal pattern in MDDs. VMHC alterations were markedly related to episode type and illness severity. The higher Hamilton Depression Rating Scale score, the more apparent VMHC reduction in the primary visual cortex. First-episode MDDs revealed stronger VMHC reduction in PCu relative to recurrent MDDs.We confirmed a significant VMHC reduction in MDDs in broad areas, especially in PCC/PCu. This reduction was affected by gender, age, episode type, and illness severity. These findings suggest that the depressive brain tends to disconnect information exchange across hemispheres.

Purpose: To establish recommended phase II dose (RP2D) in phase I and evaluate safety and efficacy of abivertinib in patients with EGFR Thr790Met point mutation (T790M)-positive(+) non–small cell lung cancer (NSCLC) with disease progression from prior EGFR inhibitors in phase II. Patients and Methods: This multicenter, open-label study included 367 adult Chinese patients. Abivertinib at doses of 50 mg twice a day to 350 mg twice a day was evaluated in phase I in continual 28-day cycles, and the RP2D of 300 mg twice a day was used in phase II in continual 21-day cycles. Primary endpoints include RP2D in phase I and objective response rate (ORR) at RP2D in phase II. Results: The RP2D of 300 mg twice a day for abivertinib was established based on pharmacokinetics, efficacy, and safety profiles across doses in phase I. In phase II, 227 patients received RP2D for a median treatment duration of 24.6 weeks (0.43–129). Among 209 response–evaluable patients, confirmed ORR was 52.2% [109/209; 95% confidence interval (CI): 45.2–59.1]. Disease control rate (DCR) was 88.0% (184/209; 95% CI: 82.9–92.1). The median duration of response (DoR) and progression-free survival (PFS) was 8.5 months (95% CI: 6.1–9.2) and 7.5 months (95% CI: 6.0–8.8), respectively. The median overall survival (OS) was 24.9 months [95% CI: 22.4–not reachable (NR)]. All (227/227) patients reported at least 1 adverse event (AE), with 96.9% (220/227) of treatment-related AEs. Treatment-related serious AEs were reported in 13.7% (31/227) of patients. Death was reported in 4.4% (10/227) of patients, and none was deemed as treatment-related. Conclusions: Abivertinib of 300 mg twice a day demonstrated favorable clinical efficacy with manageable side effects in patients with EGFR T790M+ NSCLC.

Because of paucity of published data, we evaluated the 2-year major clinical outcomes between early invasive (EI) and delayed invasive (DI) strategies according to the stage of chronic kidney disease (CKD) in patients with non-ST-segment elevation myocardial infarction (NSTEMI), who underwent a successful newer-generation drug-eluting stent (DES) implantation.A total of 8241 NSTEMI patients were recruited from the Korea Acute Myocardial Infarction Registry (KAMIR). Based on baseline estimated glomerular filtration rate (eGFR; ≥90, 60-89, 30-59, and30 mL/min/1.73 mAfter multivariable-adjusted and propensity score-adjusted analyses, the cumulative incidence of MACE (group A, p = 0.139 and p = 0.103, respectively; group B, p = 0.968 and p = 0.608, respectively; group C, p = 0.111 and p = 0.196, respectively; group D, p = 0.882 and p = 0.571, respectively), all-cause death, re-MI, and any repeat revascularization was similar between the EI and DI groups in the 4 different renal function groups.In the era of newer-generation DES, EI and DI strategies showed comparable major clinical outcomes in patients with NSTEMI and CKD during a 2-year follow-up period. However, to confirm these results, further randomized, large-scale, long-term follow-up studies are needed.

This paper presents the field programmable gate array (FPGA) implementation of a particle swarm optimization (PSO) based nonlinear model predictive control (NMPC) for applications with millisecond timescales and resource-constrained embedded systems. A custom hardware circuit architecture of NMPC controller, which mainly includes phase locking loop (PLL) clock, universal asynchronous receiver transmitter (UART) interface, trigonometric function, random number generator, objective function, PSO solver modules, is implemented on FPGA based on a hardware description language (HDL) specification. It employs parallelism, pipelining, and specialized numerical formats to enhance the online computation performance of NMPC. And the partial parallel expansion calculations of fitness (objective) function and particle position update are adopted for trading off the computational performance against resource usage. The hardware-in-the-loop experimental results through controlling a wheeled mobile robot model demonstrate that the design is capable of computing the required control action in real-time by using a midrange FPGA, while still offering good control performance.

Few-shot learning is an emerging yet challenging problem in which the goal is to achieve generalization from only few examples. Meta-learning tackles few-shot learning via the learning of prior knowledge shared across tasks and using it to learn new tasks. One of the most representative meta-learning algorithms is the model-agnostic meta-learning (MAML), which formulates prior knowledge as a common initialization, a shared starting point from where a learner can quickly adapt to unseen tasks. However, forcibly sharing an initialization can lead to conflicts among tasks and the compromised (undesired by tasks) location on optimization landscape, thereby hindering task adaptation. Furthermore, the degree of conflict is observed to vary not only among the tasks but also among the layers of a neural network. Thus, we propose task-and-layer-wise attenuation on the compromised initialization to reduce its adverse influence on task adaptation. As attenuation dynamically controls (or selectively forgets) the influence of the compromised prior knowledge for a given task and each layer, we name our method Learn to Forget (L2F). Experimental results demonstrate that the proposed method greatly improves the performance of the state-of-the-art MAML-based frameworks across diverse domains: few-shot classification, cross-domain few-shot classification, regression, reinforcement learning, and visual tracking.

Person-centred care has the potential to improve the patient experience in the intensive care unit (ICU). However, the relationship between person-centred care perceived by critically ill patients and their ICU experience has yet to be determined.The aim of this study was to investigate the relationship between person-centred care and the ICU experience of critically ill patients.This study was a multicentre, cross-sectional survey involving 19 ICUs of four university hospitals in Busan, Korea. The survey was conducted from June 2019 to July 2020, and 787 patients who had been admitted to the ICU for more than 24 hours participated. We measured person-centred care using the Person-Centered Critical Care Nursing perceived by Patient Questionnaire. Participants' ICU experience was measured by the Korean version of the Intensive Care Experience Questionnaire that consists of four subscales. We analysed the relationship between person-centred care and each area of the ICU experience using multivariate linear regression.Person-centred care was associated with 'awareness of surroundings' (β = 0.29, p lt; .001), 'frightening experiences' (β = -0.31, p lt; .001), and 'satisfaction with care' (β = 0.54, p lt; .001). However, there was no significant association between person-centred care and 'recall of experience'.We observed that person-centred care was positively related to most of the ICU experiences of critically ill patients except for recall of experience. Further studies on developing person-centred nursing interventions are needed.

Background Femoral revision surgery in patients with substantial bone loss is challenging. Impaction bone grafting using a cemented stem can be a good solution for reconstruction of the femur with poor bone stock and extensive bone loss. This study aimed to evaluate the mid-to-long-term clinical and radiographic results of impaction bone grafting using a cemented stem for Paprosky IV femoral bone defects. Methods Thirteen patients (13 hips) who underwent revision total hip arthroplasty with impaction bone grafting using a cemented stem and were followed up for at least 5 years were enrolled in this study. In all patients, a sufficient amount of fresh frozen bone of good quality was used. When cortical segmental defects were present, peripheral reinforcement with metal mesh and strut allograft was performed. The average follow-up duration was 11.1 (range, 5.3–15.1) years. The clinical and radiographic outcomes were reviewed at the final follow-up. Results The average Harris hip score was 82.5 (range, 79–94), and the average University of California, Los Angeles activity score was 5.6 (range, 4–8) at the final follow-up. Radiographic assessment revealed an average femoral component subsidence level of 0.67 (range, 0.05–2.81) mm. There were no complications, except one case (7.6%) of periprosthetic fracture. Conclusions Impaction bone grafting using a cemented stem yielded excellent mid-to-long-term outcomes. It is a reliable technique for Paprosky IV femoral bone defects, and even when severe femoral cortical bone defects are present, long-term stability can be obtained using a metal mesh and/or strut allograft.

Deep reinforcement learning (DRL)-based recommender systems have recently come into the limelight due to their ability to optimize long-term user engagement. A significant challenge in DRL-based recommender systems is the large action space required to represent a variety of items. The large action space weakens the sampling efficiency and thereby, affects the recommendation accuracy. In this article, we propose a DRL-based method called deep hierarchical category-based recommender system (DHCRS) to handle the large action space problem. In DHCRS, categories of items are used to reconstruct the original flat action space into a two-level category-item hierarchy. DHCRS uses two deep Q -networks (DQNs): 1) a high-level DQN for selecting a category and 2) a low-level DQN to choose an item in this category for the recommendation. Hence, the action space of each DQN is significantly reduced. Furthermore, the categorization of items helps capture the users' preferences more effectively. We also propose a bidirectional category selection (BCS) technique, which explicitly considers the category-item relationships. The experiments show that DHCRS can significantly outperform state-of-the-art methods in terms of hit rate and normalized discounted cumulative gain for long-term recommendations.

Objectives To assess treatment outcome (transversal and sagittal dental arch relationships) and its determinants in complete bilateral cleft lip and palate (BCLP) evaluated with the modified Huddart-Bodenham scoring system and the BCLP Yardstick. Materials and methods Multiple electronic databases were searched without time limitation. Randomized clinical trials, cohort and case control studies using BCLP Yardstick and/or modified Huddart-Bodenham system to judge treatment outcome of patients with BCLP were included. The Risk of Bias in Nonrandomized Studies of Interventions tool and Grading of Recommendations, Assessment, Development, and Evaluation was used. Results Of the 528 studies identified by the electronic search, only eight retrospective studies met the inclusion criteria and were included. A total of 12 cleft centers were represented. All treatment protocols differed and background information was underreported. The results for the BCLP yardstick showed that all except the centers in New Zealand had a mean score lower than 3, indicating good treatment results. However, these studies had a moderate to high risk of bias. The modified Huddart-Bodenham scores were negative in all studies. No further meta-analysis was done due to heterogeneity and high risk of bias. The quality of evidence was graded as very low. Conclusion Results for the dental arch relationship of studies in complete BCLP and possible determinants were not synthesized due to very low quality of evidence. Clinical research for patients with BCLP should focus on sound methodological designs to enable evidence-based decision making to improve treatment for patients with BCLP and thereby hopefully their quality of life.

Anti-PD-1 and oncolytic viruses (OVs) have non-overlapping anti-tumor mechanisms, since each agent works at different steps of the cancer-immunity cycle. Evidence suggests that OVs improve therapeutic responses to anti-PD-1 therapy by reversing immunosuppressive factors, increasing the number and diversity of infiltrating lymphocytes, and promoting PD-L1 expression in both injected and non-injected tumors. Many studies in preclinical models suggest that the timing of anti-PD-1 administration influences the therapeutic success of the combination therapy (anti-PD-1 + OV). Therefore, determining the appropriate sequencing of agents is of critical importance to designing a rationale OV-based combinational clinical trial. Currently, the combination of anti-PD-1 and OVs are being delivered using various schedules, and we have classified the timing of administration of anti-PD-1 and OVs into five categories: (i) anti-PD-1 lead-in → OV; (ii) concurrent administration; (iii) OV lead-in → anti-PD-1; (iv) concurrent therapy lead-in → anti-PD-1; and (v) OV lead-in → concurrent therapy. Based on the reported preclinical and clinical literature, the most promising treatment strategy to date is hypothesized to be OV lead-in → concurrent therapy. In the OV lead-in → concurrent therapy approach, initial OV treatment results in T cell priming and infiltration into tumors and an immunologically hot tumor microenvironment (TME), which can be counterbalanced by engagement of PD-L1 to PD-1 receptor on immune cells, leading to T cell exhaustion. Therefore, after initial OV therapy, concurrent use of both OV and anti-PD-1 is critical through which OV maintains T cell priming and an immunologically hot TME, whereas PD-1 blockade helps to overcome PD-L1/PD-1-mediated T cell exhaustion. It is important to note that the hypothetical conclusion drawn in this review is based on thorough literature review on current understanding of OV + anti-PD-1 combination therapies and rhythm of treatment-induced cancer-immunity cycle. A variety of confounding factors such as tumor types, OV types, presence or absence of cytokine transgenes carried by an OV, timing of treatment initiation, varying dosages and treatment frequencies/duration of OV and anti-PD-1, etc. may affect the validity of our conclusion that will need to be further examined by future research (such as side-by-side comparative studies using all five treatment schedules in a given tumor model).

Ground military target recognition plays a crucial role in unmanned equipment and grasping the battlefield dynamics for military applications, but is disturbed by low-resolution and noisy-representation. In this paper, a recognition method, involving a novel visual attention mechanism-based Gabor region proposal sub-network (Gabor RPN) and improved refinement generative adversarial sub-network (GAN), is proposed. Novel central–peripheral rivalry 3D color Gabor filters are proposed to simulate retinal structures and taken as feature extraction convolutional kernels in low-level layer to improve the recognition accuracy and framework training efficiency in Gabor RPN. Improved refinement GAN is used to solve the problem of blurry target classification, involving a generator to directly generate large high-resolution images from small blurry ones and a discriminator to distinguish not only real images vs. fake images but also the class of targets. A special recognition dataset for ground military target, named Ground Military Target Dataset (GMTD), is constructed. Experiments performed on the GMTD dataset effectively demonstrate that our method can achieve better energy-saving and recognition results when low-resolution and noisy-representation targets are involved, thus ensuring this algorithm a good engineering application prospect.

With the maturity and popularity of Internet of Things (IoT), the notion of Social Internet of Things (SIoT) has been proposed to support novel applications and networking services for the IoT in more effective and efficient ways. In this paper, we propose a new problem family, namely, Task-Optimized SIoT Selection (TOSS), to find the best group of IoT objects for a given set of tasks in the task pool. TOSS aims to select the target SIoT group such that the target SIoT group is able to easily communicate with each other while maximizing the accuracy of performing the given tasks. We propose two problem formulations, named Bounded Communication-loss TOSS (BC-TOSS) and Robustness Guaranteed TOSS (RG-TOSS), for different scenarios and prove that they are both NP-hard and inapproximable. We propose a polynomial-time algorithm with a performance guarantee for BC-TOSS, and an efficient polynomial-time algorithm to obtain good solutions for RG-TOSS. Moreover, as RG-TOSS is NP-hard and inapproximable within any factor, we further propose Structure-Aware Reinforcement Learning (SARL) to leverage the Graph Convolutional Networks (GCN) and Deep Reinforcement Learning (DRL) to effectively solve RG-TOSS. Experimental results indicate that our proposed algorithms outperform the other deterministic and learning-based baseline approaches.

Deep neural networks have achieved breakthrough improvement in various application fields. Nevertheless, they usually suffer from a time-consuming training process because of the complicated structures of neural networks with a huge number of parameters. As an alternative, a fast and efficient discriminative broad learning system (BLS) is proposed, which takes the advantages of flat structure and incremental learning. The BLS has achieved outstanding performance in classification and regression problems. However, the previous studies ignored the reason why the BLS can generalize well. In this article, we focus on the interpretation from the viewpoint of the frequency domain. We discover the existence of the frequency principle in BLS, i.e., the BLS preferentially captures low-frequency components quickly and then fits the high frequencies during the incremental process of adding feature nodes and enhancement nodes. The frequency principle may be of great inspiration for expanding the application of BLS.

In the current era of the Internet of Things (IoT), various devices can provide more services by connecting to the Internet. However, the explosive growth of connected devices will cause the cloud core overload and significant network delays. To overcome these problems, the Mobile Edge Computing (MEC) network is proposed to provide most of the computing and storage near the radio access network to reduce the traffic of the core cloud network and provide lower latency for the terminal.Mobile edge computing can work with third parties to develop multiple services, such as mobile big data analysis and context-aware services. However, when there is a large amount of popular data accessed in a short period, the system must generate many replicas, which will not only reduce access efficiency but also cause additional traffic overhead. To improve the above problems, an Adaptive Replica Configuration Mechanism (ARCM) is proposed in this paper to predict the popularity of the file and make a replica to the low-blocking node. This method spreads the subsequent access workload by copying the popular file in advance to improve the overall performance of the system.

A dilemma faced by classification is that the data is not collected at the same frequency in some applications. We investigate the mixed frequency data in a new way and recognize them as a special style of multi-view data, in which each view data is collected at a different sampling frequency. This paper proposes a discriminative dictionary pair learning method constrained by ordinal locality for mixed frequency data classification (shorted by DPLOL-MF). This method integrates synthesis dictionary and analysis dictionary into a dictionary pair, which not only improves computational cost caused by the ${\ell_0}$ or ${\ell_1}$ -norm constraint, but also can deal with the sampling frequency inconsistency. The DPLOL-MF utilizes a synthesis dictionary to learn class-specified reconstruction information and employs an analysis dictionary to generate coding coefficients by analyzing samples. Particularly, the ordinal locality preserving term is leveraged to constrain the atoms of dictionaries pair to further facilitate the learned dictionary pair to be more discriminative. Besides, we design a specific classification scheme for the inconsistent sample size of mixed frequency data. This paper illustrates a novel idea to solve the classification task of mixed frequency data and the experimental results demonstrate the effectiveness of the proposed method.

Scene classification of high spatial resolution (HSR) images can provide data support for many practical applications, such as land planning and utilization, and it has been a crucial research topic in the remote sensing (RS) community. Recently, deep learning methods driven by massive data show the impressive ability of feature learning in the field of HSR scene classification, especially convolutional neural networks (CNNs). Although traditional CNNs achieve good classification results, it is difficult for them to effectively capture potential context relationships. The graphs have powerful capacity to represent the relevance of data, and graph-based deep learning methods can spontaneously learn intrinsic attributes contained in RS images. Inspired by the abovementioned facts, we develop a deep feature aggregation framework driven by graph convolutional network (DFAGCN) for the HSR scene classification. First, the off-the-shelf CNN pretrained on ImageNet is employed to obtain multilayer features. Second, a graph convolutional network-based model is introduced to effectively reveal patch-to-patch correlations of convolutional feature maps, and more refined features can be harvested. Finally, a weighted concatenation method is adopted to integrate multiple features (i.e., multilayer convolutional features and fully connected features) by introducing three weighting coefficients, and then a linear classifier is employed to predict semantic classes of query images. Experimental results performed on the UCM, AID, RSSCN7, and NWPU-RESISC45 data sets demonstrate that the proposed DFAGCN framework obtains more competitive performance than some state-of-the-art methods of scene classification in terms of OAs.