Search

Adversarial attacks have been extensively studied in recent years since they can identify the vulnerability of deep learning models before deployed. In this paper, we consider the black-box adversarial setting, where the adversary needs to craft adversarial examples without access to the gradients of a target model. Previous methods attempted to approximate the true gradient either by using the transfer gradient of a surrogate white-box model or based on the feedback of model queries. However, the existing methods inevitably suffer from low attack success rates or poor query efficiency since it is difficult to estimate the gradient in a high-dimensional input space with limited information. To address these problems and improve black-box attacks, we propose two prior-guided random gradient-free (PRGF) algorithms based on biased sampling and gradient averaging, respectively. Our methods can take the advantage of a transfer-based prior given by the gradient of a surrogate model and the query information simultaneously. Through theoretical analyses, the transfer-based prior is appropriately integrated with model queries by an optimal coefficient in each method. Extensive experiments demonstrate that, in comparison with the alternative state-of-the-arts, both of our methods require much fewer queries to attack black-box models with higher success rates., Accepted by IEEE Transactions on Pattern Recognition and Machine Intelligence (TPAMI). The official version is at https://ieeexplore.ieee.org/document/9609659

Leying Xi,1 Shuang Xu,2 Yingying Jiang,2 Hang Su,2 Yuying Sun,2 Yingying Wen,2 Jingjing Wu,2 Xianqing Ren2 1Department of Pediatrics, Nanjing Hospital of Chinese Medicine Affiliated to Nanjing University of Chinese Medicine, Nanjing, Jiangsu, People’s Republic of China; 2Department of Pediatrics, The First Affiliated Hospital of Henan University of Chinese Medicine, Zhengzhou, Henan, People’s Republic of ChinaCorrespondence: Xianqing RenDepartment of Pediatrics, The First Affiliated Hospital of Henan University of Chinese Medicine, No. 9 Renmin Road, Jinshui District, Zhengzhou, Henan Province, 450000, People’s Republic of ChinaTel +86 13673712698Email Renxq723@163.comObjective: Henoch–Schönlein purpura (HSP) is the most common vasculitis in children. Renal involvement is the main long-term complication of HSP, and presently there is no way to predict which children may have irreversible renal damage from the outset. This study aimed to explore the prediction value of laboratory indexes on renal involvement in children with HSP, which could help the early identification and intervention.Methods: Children with HSP hospitalized at the First Affiliated Hospital of Henan University of Chinese Medicine from June 2019 to December 2020 were included. The demographic findings, clinical features, laboratory findings including blood routine examination, serum immunoglobulin, complement, T cell subsets levels, liver and kidney function, coagulation function were recorded. Laboratory indexes were analyzed, logistic regression analysis was performed to identify the independent predictors in HSP patients with renal involvement, and receiver operating characteristic (ROC) curves were further used to assess the value of prediction indexes, as well as the efficacy of combined diagnosis.Results: The study included 146 HSP patients, among them, 50 patients (34.2%) had renal involvement. Age, platelet distribution width (PDW), CD3+ and fibrinogen (FIB) were positively correlated with renal involvement, while the levels of Immunoglobulin G (IgG), C-reactive protein (CRP), and neutrophil/lymphocyte ratio (NLR) were negatively correlated with renal involvement. The area under the ROC Curve (AUC) of these biomarkers ranged from 0.6284 to 0.7009, and among the combinations, a combination of NLR, CRP, CD3+, FIB, PDW, IgG and age had the best AUC value (0.9774).Conclusion: Age, PDW, CD3+, FIB, CRP, NLR and IgG were prediction indexes for renal involvement in HSP patients, and these indexes can be combined appropriately to improve the diagnostic efficacy.Keywords: Henoch–Schönlein purpura, laboratory indexes, renal involvement, diagnosis

Objective: In order to examine the effect of 4-week combinatorial intervention with exercise training and dietary restriction on serum fatty acids, and to explore the correlation of intervention-induced improvement of serum fatty acid profile with the reduction of body fat and arterial stiffness. Methods: Thirty-three obese adolescents were randomized into the intervention (n = 19) and control (n = 14) groups. The participants from the intervention group were subjected to 4-week combinatorial intervention with exercise training and dietary restriction while the participants from the control group maintained regular activities and diet habits. Anthropometry, serum fatty acids and arterial stiffness were measured before and after 4-week intervention. Results: The participants upon combinatorial intervention revealed the improved body compositions and serum fatty acid profile, and reduced arterial stiffness when compared with their basal levels and the control participants (p < 0.05). Moreover, the decrease in myristic acid, stearic acid, arachidic acid, behenic acid, palmitoleic acid, and dihomo-γ-linolenic acid, was associated with the reduction in body fat. A positive correlation between arachidonic acid and left brachial ankle pulse velocity was observed, and the increase in docosahexaenoic acid was associated with the reduction of left brachial ankle pulse wave velocity and the enhancement of right ankle brachial index. Conclusion: The 4-week combinatorial intervention is a useful strategy to improve serum fatty acid profile along with the reduction of body fat and arterial stiffness in obese adolescents.

BACKGROUND: Inflammation is one of major contributors of diabetic osteoporosis. Here, we combined adipose tissues derived mesenchymal stem cells (AD-MSCs)-derived exosomes and microRNA-146a (miR-146a) to develop more effective anti-inflammation strategy in osteoclasts. METHODS: miR-146a was overexpressed in AD-MSCs and miR-146a exosomes (miR-146a-Exo) were isolated and characterized. Cellular and animal diabetic osteoporosis models were created to evaluate the anti-inflammation effect of miR-146a-Exo by using ELISA, qRT-PCR, MTT, bone resorption assay, Western blot, and bone mineral content and density analysis in vitro and in vivo. RESULTS: miR-146a-Exo administration presented the most potent effect on inhibition of pro-inflammatory cytokines production in high glucose-treated osteoclasts, restraint bone resorption, and restoration of the bone loss in streptozotocin-induced diabetic osteoporosis rats. Mechanistically, miR-146a-Exo suppressed the expression of TNF-α, IL-18, and IL-1β, induced the inactivation of inflammasome, and finally reduced bone resorption and recovered bone loss. CONCLUSION: Combination of AD-MSCs-Exo and miR-146a more effectively exert the anti-inflammation effect in osteoclasts, providing a potential drug for the treatment of diabetic osteoporosis.

Objective: To assess the efficacy and safety of the novel histone deacetylase inhibitor, chidamide, in combination with cyclophosphamide, doxorubicin, vincristine, etoposide, and prednisone (Chi-CHOEP) for untreated peripheral T-cell lymphoma (PTCL). Methods: A prospective, multicenter, single arm, phase 1b/2 study was conducted. A total of 128 patients with untreated PTCL (18–70 years of age) were enrolled between March 2016 and November 2019, and treated with up to 6 cycles with the Chi-CHOEP regimen. In the phase 1b study, 3 dose levels of chidamide were evaluated and the primary endpoint was determination of the maximum-tolerated dose and recommended phase 2 dose (RP2D). The primary endpoint of the phase 2 study was 2-year progression-free survival (PFS). Results: Fifteen patients were enrolled in the phase 1b study and the RP2D for chidamide was determined to be 20 mg, twice a week. A total of 113 patients were treated at the RP2D in the phase 2 study, and the overall response rate was 60.2%, with a complete response rate of 40.7%. At a median follow-up of 36 months, the median PFS was 10.7 months, with 1-, 2-, and 3-year PFS rates of 49.9%, 38.0%, and 32.8%, respectively. The Chi-CHOEP regimen was well-tolerated, with grade 3/4 neutropenia occurring in approximately two-thirds of the patients. No unexpected adverse events (AEs) were reported and the observed AEs were manageable. Conclusions: This large cohort phase 1b/2 study showed that Chi-CHOEP was well-tolerated with modest efficacy in previously untreated PTCL patients.

Background The eighth edition of nodal classification for non-small cell lung cancer (NSCLC) is defined only by the anatomical location of metastatic lymph nodes. Research Question We sought to evaluate the prognostic significance and discriminatory capability of the number of involved nodal stations (nS) in a large Chinese cohort. Study Design and Methods A total of 4,011 patients with NSCLC undergoing surgical resection between 2009 and 2013 were identified. The optimal cutoff values for nS classification were determined with X-tile software. Kaplan-Meier and multivariate Cox analysis were used to examine the prognostic performance of nS classification in comparison with location-based N classification. A decision curve analysis was performed to evaluate the standardized net benefit of nS classification in predicting prognosis. Results All the patients were classified into four prognostically different subgroups according to the number of involved nodal stations: (1) nS0 (none positive), (2) nS1 (one involved station), (3) nS2 (two involved stations), and (4) nS ≥ 3 (three or more involved stations). The prognoses among all the neighboring categories of nS classification were statistically significantly different in terms of disease-free survival and overall survival. The multivariate Cox analysis demonstrated that nS was an independent prognostic factor of disease-free survival and overall survival. Patients with N1 or N2 stage disease could be divided into three prognostically different subgroups according to nS classification. However, the prognosis was similar between the N1 and N2 subgroups when patients were staged in the same nS category. The decision curve analysis showed that nS classification tended to have a higher predictive capability than location-based N classification. Interpretation The nS classification could be used to provide a more accurate prognosis for patients with resected NSCLC. The nS is worth taking into consideration when defining nodal category in the forthcoming ninth edition of the staging system.

Inflammation is one of major contributors of diabetic osteoporosis. Adipose derived mesenchymal stem cells (AD-MSCs) show great potential to inhibit inflammation. We investigated the anti-osteoporosis role of AD-MSCs-derived exosomes in diabetic osteoporosis and the underlying molecular mechanism. Cellular and animal diabetic osteoporosis models were created through high glucose exposure and streptozotocin injection. AD-MSCs-derived exosomes were isolated and characterized. Pro-inflammatory cytokines and osteoclast markers were determined by ELISA. Bone mineral content and density were detected to evaluate bone loss. qRT-PCR and Western blots were performed to detect the expression of target genes. AD-MSCs-derived exosomes inhibited the secretion of IL-1β and IL-18 in HG treated osteoclasts and restored the bone loss in streptozotocin-induced diabetic osteoporosis rats. Mechanistically, AD-MSCs-derived exosomes suppress NLRP3 inflammasome activation in osteoclasts, and then reduce bone resorption and recover bone loss. AD-MSCs-derived exosomes alleviate diabetic osteoporosis through suppressing NLRP3 inflammasome activation in osteoclasts, which might be a potential cell-free therapeutic approach for diabetes-induced bone loss treatment.

Purpose Pelvic fractures are characterized by high energy injuries and often accompanied with abdominal and pelvic organ injury. CT has been applied for several decades to evaluate blunt pelvic trauma patients. However, it has a certain rate of inaccurate diagnosis of abdominal hollow viscus injury (HVI), especially in the early stage after injury. The delayed diagnosis of HVI could result in a high morbidity and mortality. The bowel injury prediction score (BIPS) applied 3 clinical variables to determine whether an early surgical intervention for blunt HVI was necessary. We recently found another clinical variable (iliac ecchymosis, IE) which appeared at the early stage of injury, could be predicted for HVI. The main objective of this study was to explore the novel combination of IE and BIPS to enhance the early diagnosis rate of HVI, and thus reduce complications and mortalities. Methods We conducted a retrospective analysis from January 2008 to December 2018 and recorded blunt pelvic trauma patients in our hospital. The inclusion criteria were patients who were verified with pelvic fractures using abdomen and pelvis CT scan in the emergency department before any surgical intervention. The exclusion criteria were abdominal CT insufficiency before operation, abdominal surgery before CT scan, and CT mesenteric injury grade being 5. The MBIPS was defined as BIPS plus IE, which was calculated according to 4 variables: white blood cell counts of 17.0 or greater, abdominal tenderness, CT scan grade for mesenteric injury of 4 or higher, and the location of IE. Each clinical variable counted 1 score, totally 4 scores. The location and severity of IE was also noted. Results In total, 635 cases were hospitalized and 62 patients were enrolled in this study. Of these included patients, 77.4% (40 males and 8 females) were operated by exploratory laparotomy and 22.6% (8 males and 6 females) were treated conservatively. In the 48 patients underwent surgical intervention, 46 were confirmed with HVI (45 with IE and 1 without IE). In 46 patients confirmed without HVI, only 3 patients had IE and the rest had no IE. The sensitivity and specificity of IE in predicting HVI was calculated as 97.8% (45/46) and 81.3% (13/16), respectively. The median MBIPS score for surgery group was 2, while 0 for the conservative treatment group. The incidence of HVI in patients with MBIPS score ≥ 2 was significantly higher than that in patients with MBIPS score less than ≤ 2 (OR = 17.3, p Conclusion IE can be recognized as an indirect sign of HVI because of the high sensitivity and specificity, which is a valuable sign for HVI in blunt pelvic trauma patients. MBIPS can be used to predict HVI in blunt pelvic trauma patients. When the MBIPS score is ≥ 2, HVI is strongly suggested.

Our study aimed to validate the clinicopathological characteristics and prognosis of lung adenocarcinoma (ADC) with a filigree pattern and to further investigate the relationship between the filigree pattern and the classical micropapillary (MP) pattern. We retrospectively reviewed the clinical and pathologic characteristics of 461 Chinese patients with completely resected ADC (stage I, 310; stage II, 44; stage III, 107). The filigree pattern was more likely to be observed in ADC with a higher stage (p = 0.003) and the classical MP pattern (p < 0.001). Patients with filigree-predominant ADC showed poor survival, similar to those with classical MP-predominant ADC. Multivariate analysis confirmed that the presence of the filigree pattern was an independent prognostic factor for recurrence-free survival (hazard ratio (HR), 2.01; 95% confidence interval (CI), 1.50-2.68; p < 0.001) and overall survival (OS; HR, 1.83; 95% CI, 1.34-2.50; p < 0.001). Patients with both classical MP-positive and filigree-positive ADC had the worst survival compared with those with the filigree pattern or classical MP pattern alone. In stage I, ADC with both the filigree and classical MP patterns had a higher incidence of micrometastasis than ADC with the filigree pattern or classical MP pattern alone. Lymph node micrometastasis indicated poor survival in patients with ADC with the filigree pattern or classical MP pattern. Similar clinicopathologic features between patients with the filigree pattern and the classical MP pattern support the inclusion of the filigree pattern in the MP category. Recognition of the filigree pattern could provide helpful prognostic information, especially for stage I ADC.

Enzymatic biofuel cells (EBFCs) have great application prospects by catalyzing the redox reaction of biofuels to generate electricity. In order to further improve the output performance of EBFCs, the influences of electrode shape and array on the outputs of EBFCs were investigated by finite element analysis. The relationship between electrode shapes and the distribution of field strength and voltage were firstly explored by graphical method. According to simulation results, circular electrodes distribution would be the best choice for EBFCs with a maximum power density (MPD) of over 4.7 mW cm−2 and the least current density range of 4.4 mA cm−2 under ideal conditions. The hexagonal close-packed array was able to distribute more electrodes and thus obtain higher total power outputs. For the monopolar plate EBFCs, anodes and cathodes were interlaced on the same plate to facilitate the application of wearable and implantable devices. Moreover, the network topology theory was applied to the design of electrode array in EBFCs, which provided extensive electrode array structures for EBFCs. The results exhibited that the MPD of radial and star topologies were both about 4.7 mW cm−2, while that of ring topology was much smaller. The electrode shape and electrode array studied in this paper would be of great significance to the structural design of EBFCs devices.

Purpose The purpose of this paper is to mainly center on the touchless interaction between humans and robots in the real world. The accuracy of hand pose identification and stable operation in a non-stationary environment is the main challenge, especially in multiple sensors conditions. To guarantee the human-machine interaction system’s performance with a high recognition rate and lower computational time, an adaptive sensor fusion labeling framework should be considered in surgery robot teleoperation. Design/methodology/approach In this paper, a hand pose estimation model is proposed consisting of automatic labeling and classified based on a deep convolutional neural networks (DCNN) structure. Subsequently, an adaptive sensor fusion methodology is proposed for hand pose estimation with two leap motions. The sensor fusion system is implemented to process depth data and electromyography signals capturing from Myo Armband and leap motion, respectively. The developed adaptive methodology can perform stable and continuous hand position estimation even when a single sensor is unable to detect a hand. Findings The proposed adaptive sensor fusion method is verified with various experiments in six degrees of freedom in space. The results showed that the clustering model acquires the highest clustering accuracy (96.31%) than other methods, which can be regarded as real gestures. Moreover, the DCNN classifier gets the highest performance (88.47% accuracy and lowest computational time) than other methods. Originality/value This study can provide theoretical and engineering guidance for hand pose recognition in surgery robot teleoperation and design a new deep learning model for accuracy enhancement.

Methamphetamine (METH), a highly addictive psychostimulant, is the second most widely used illicit drug. METH produces damage dopamine neurons and apoptosis via multiple inter-regulating mechanisms, including dopamine overload, hyperthermia, oxidative stress, mitochondria dysfunction, endoplasmic reticulum stress, protein degradation system dysfunction, and neuroinflammation. Increasing evidence suggests that chronic METH abuse is associated with neurodegenerative changes in the human brain and an increased risk of Parkinson's disease (PD). METH use and PD may share some common steps in causing neurotoxicity. Accumulation of α-synuclein, a presynaptic protein, is the pathological hallmark of PD. Intriguingly, α-synuclein upregulation and aggregation are also found in dopaminergic neurons in the substantia nigra in chronic METH users. This suggests α-synuclein may play a role in METH-induced neurotoxicity. The mechanism of α-synuclein cytotoxicity in PD has attracted considerable attention; however, how α-synuclein affects METH-induced neurotoxicity has not been reviewed. In this review, we summarize the relationship between METH use and PD, interdependent mechanisms that are involved in METH-induced neurotoxicity and the significance of α-synuclein upregulation in response to METH use. The identification of α-synuclein overexpression and aggregation as a contributor to METH-induced neurotoxicity may provide a novel therapeutic target for the treatment of the deleterious effect of this drug and drug addiction.

Background Craving and cognitive deficits are potential treatment targets for methamphetamine use disorder (MUD). Previous studies implied that transcranial direct current stimulation (tDCS) and cognitive training respectively improve these symptoms, but the combined effect is unknown. In this study, we investigated the combined effects of tDCS over dorsolateral prefrontal cortex (DLPFC) and computerized cognitive addiction therapy (CCAT) on cue-induced craving and cognitive functions among female individuals with MUD. Methods Seventy-five patients with MUD were randomly assigned to three groups: CCAT + tDCS group, CCAT + sham tDCS group and the control group. The former two groups received 20 sessions of cognitive training combined 1.5 mA active/sham tDCS over DLPFC (20min/session, 5times/week), while the control group received usual care which includes routine medical care, health education, physical exercises and psychological support related to relapse prevention. The cue-induced craving and cognitive functions were tested at the baseline, the end of 2nd week and 4th week. Results The CCAT + tDCS group showed a significant reduction in cue-induced craving after 4-week intervention. Moreover, the craving score of the real CCAT + tDCS group was significantly lower than that of the CCAT + sham tDCS group and that of the control group at the end of 4th week. A significant improvement in accuracy of TWOB task was only observed in the CCAT + tDCS group at the end of 4th week when compared to baseline. Unexpectedly, participants who received CCAT plus active or sham tDCS did not change their discounting, whereas those in the control group performed more impulsively over time. Conclusions The study found that the intervention of tDCS over DLPFC combined with CCAT may have potential benefit in improving treatment outcome in patients with MUD. More research is needed to explore the underlying mechanism.

Background Invasive adenocarcinoma intraoperatively underestimated as adenocarcinoma in situ (AIS) or minimally invasive adenocarcinoma (MIA) based on frozen section (FS) is more likely to undergo insufficient resection. We aimed to investigate the predictors of upstage and treatment strategies for stage IA invasive adenocarcinoma after sublobar resection for AIS and MIA. Methods We identified 2,006 patients from January 2012 to December 2016 with early-stage lung adenocarcinoma who underwent sublobar resection based on FS diagnosis to guide surgical decision-making. All FS were categorized into three groups in real-time: (I) atypical adenomatous hyperplasia (AAH), (II) AIS, and (III) MIA. Results A total of 272 (13.5%, 272/2,006) cases were upstaged in the final pathology (FP) diagnosis (82 AAH to AIS, 127 AIS to MIA, and nine AIS and 54 MIA to invasive adenocarcinoma), and most upstage cases (64.3%, 175/272) were attributed to sampling error. Multivariate logistic regression showed that tumor size ≥1 cm was the only independent predictor of upstage. The upstage of 209 cases to AIS or MIA had no influence on the therapy because the extent of their resection was enough. Of the 63 cases upstaged to invasive adenocarcinoma, only 13 cases agreed to receive complementary treatment: five patients underwent complementary lobectomy, and seven patients received chemotherapy. Two invasive adenocarcinoma cases without complementary treatment experienced a local recurrence after surgery. No recurrence was observed in AAH, AIS and MIA. No patient died until December 01, 2019. Conclusions Timely complementary treatment is encouraged in AIS/MIA upstaged to invasive adenocarcinoma based on the FP after sublobar resection to avoid local recurrence. Pathologists should be more cautious about AIS and MIA with tumor size ≥1 cm to avoid underestimation and potentially insufficient resection.

Angiogenesis contributes to numerous pathological conditions. Understanding the molecular mechanisms of angiogenesis will offer new therapeutic opportunities. Several experimental in vivo models that better represent the pathological conditions have been generated for this purpose in mice, but it is difficult to translate results from mouse to human blood vessels. To understand human vascular biology and translate findings into human research, we need human blood vessel models to replicate human vascular physiology. Here, we show that human tumor tissue transplantation into a cranial window enables engraftment of human blood vessels in mice. An in vivo imaging technique using two-photon microscopy allows continuous observation of human blood vessels until at least 49 days after tumor transplantation. These human blood vessels make connections with mouse blood vessels as shown by the finding that lectin injected into the mouse tail vein reaches the human blood vessels. Finally, this model revealed that formation and/or maintenance of human blood vessels depends on VEGFR2 signaling. This approach represents a useful tool to study molecular mechanisms of human blood vessel formation and to test effects of drugs that target human blood vessels in vivo to show proof of concept in a preclinical model.

Background Peripheral T cell lymphoma (PTCL) is a rare disease and recent approved drugs for relapsed/refractory (r/r) PTCL provided limited clinical benefit. We conducted this study to evaluate the efficacy and safety of geptanolimab (GB226), an anti-PD-1 antibody, in r/r PTCL patients. Methods We did this single-arm, multicenter phase 2 study across 41 sites in China. Eligible patients with r/r PTCL received geptanolimab 3 mg/kg intravenously every 2 weeks until disease progression or intolerable toxicity. All patients who received at least one dose of geptanolimab and histological confirmed PTCL entered full analysis set (FAS). The primary endpoint was objective response rate (ORR) in FAS assessed by the independent radiological review committee (IRRC) per Lugano 2014 criteria. Results Between July 12, 2018, and August 15, 2019, 102 patients were enrolled and received at least one dose of geptanolimab. At the data cutoff date (August 15, 2020), the median follow-up was 4.06 (range 0.30–22.9) months. For 89 patients in FAS, 36 achieved objective response (40.4%, 95% CI 30.2–51.4), of which 13 (14.6%) were complete response and 23 (25.8%) had partial response assessed by IRRC. The median duration of response (DOR) was 11.4 (95% CI 4.8 to not reached) months per IRRC. Patients with PD-L1 expression ≥ 50% derived more benefit from geptanolimab treatment compared to p = 0.013; median PFS 6.2 vs. 1.5 months, p = 0.002). Grade ≥ 3 treatment-related adverse events occurred in 26 (25.5%) patients, and the most commonly observed were lymphocyte count decreased (n = 4) and platelet count decreased (n = 3). Serious adverse events were observed in 45 (44.1%) patients and 19 (18.6%) were treatment related. Conclusions In this study, geptanolimab showed promising activity and manageable safety profile in patients with r/r PTCL. Anti-PD-1 antibody could be a new treatment approach for this patient population. Trial registration: This clinical trial was registered at the ClinicalTrials.gov (NCT03502629) on April 18, 2018.

Spatial misalignment caused by variations in poses and viewpoints is one of the most critical issues that hinders the performance improvement in existing person re-identification (Re-ID) algorithms. To address this problem, in this paper, we present a robust and efficient graph correspondence transfer (REGCT) approach for explicit spatial alignment in Re-ID. Specifically, we propose to establish the patch-wise correspondences of positive training pairs via graph matching. By exploiting both spatial and visual contexts of human appearance in graph matching, meaningful semantic correspondences can be obtained. To circumvent the cumbersome \emph{on-line} graph matching in testing phase, we propose to transfer the \emph{off-line} learned patch-wise correspondences from the positive training pairs to test pairs. In detail, for each test pair, the training pairs with similar pose-pair configurations are selected as references. The matching patterns (i.e., the correspondences) of the selected references are then utilized to calculate the patch-wise feature distances of this test pair. To enhance the robustness of correspondence transfer, we design a novel pose context descriptor to accurately model human body configurations, and present an approach to measure the similarity between a pair of pose context descriptors. Meanwhile, to improve testing efficiency, we propose a correspondence template ensemble method using the voting mechanism, which significantly reduces the amount of patch-wise matchings involved in distance calculation. With aforementioned strategies, the REGCT model can effectively and efficiently handle the spatial misalignment problem in Re-ID. Extensive experiments on five challenging benchmarks, including VIPeR, Road, PRID450S, 3DPES and CUHK01, evidence the superior performance of REGCT over other state-of-the-art approaches., Tech. Report. The source code is available at http://www.dabi.temple.edu/~hbling/code/gct.htm. arXiv admin note: text overlap with arXiv:1804.00242

Compared with stock market, cryptocurrency market is more susceptible to investor sentiment at the lack of substantial asset support. This study develops a proxy to measure the investor sentiment of cryptocurrency market by using textual analytics on millions of posts in Chain Node, which is the most active online community for Chinese cryptocurrency investors. We investigate the correlation between the sentiment and the market return from Jan. 2018 to Aug. 2020. The study argues that the proposed proxy could well reflect the investor sentiment of the cryptocurrency.

Compared with traditional centralized control strategies, the distributed control systems significantly improve the flexibility and expandability of an modular multilevel converter (MMC). However, the stability issue in the MMC distributed control system with the presence of control loop coupling interactions is rarely discussed in existing research works. This article is to improve the stability of an MMC distributed control system by inhibiting the control conflict due to the coupling interactions among control loops with incomplete control information. By modeling the MMC distributed control system, the control loop coupling interactions are analyzed and the essential cause of control conflict is revealed. Accordingly, a control parameter design principle is proposed to effectively suppress the disturbances from the targeted control conflict and improve the MMC system stability. The rationality of the theoretical analysis and the effectiveness of the control parameter design principle are confirmed by simulation and experimental results.

Objectives There is still some dispute regarding the performance of limited mediastinal lymphadenectomy (LML) even for lung adenocarcinoma ≤ 2 cm. We aimed to recognize the potential candidates who can benefit from LML based on the percentage of histological components (PHC). Methods We analyzed 1160 surgical patients with invasive lung adenocarcinoma ≤ 2 cm from seven institutions between January 2012 and December 2015. All histological subtypes were listed in 5% increments by pathological slices. To test the accuracy of frozen section in judging PHC, frozen section slides from 140 cases were reviewed by three pathologists. Results There were 882 patients with systematic mediastinal lymphadenectomy (SML) and 278 with LML. Multivariable analysis indicated that, the total percentage of micropapillary and solid components (PHCMIP+S) > 5 % was the independent predictor of N2 metastasis (P 5 % (RFS, HR = 2.143, P Conclusions PHCMIP+S showed the predictive value for N2 metastasis and procedure-specific outcome (LML vs. SML). It may serve as a feasible indicator for identifying proper candidates of LML by using intraoperative frozen section.

A comprehensive and comprehensible summary of existing deep neural networks (DNNs) helps practitioners understand the behaviour and evolution of DNNs, offers insights for architecture optimization, and sheds light on the working mechanisms of DNNs. However, this summary is hard to obtain because of the complexity and diversity of DNN architectures. To address this issue, we develop DNN Genealogy, an interactive visualization tool, to offer a visual summary of representative DNNs and their evolutionary relationships. DNN Genealogy enables users to learn DNNs from multiple aspects, including architecture, performance, and evolutionary relationships. Central to this tool is a systematic analysis and visualization of 66 representative DNNs based on our analysis of 140 papers. A directed acyclic graph is used to illustrate the evolutionary relationships among these DNNs and highlight the representative DNNs. A focus + context visualization is developed to orient users during their exploration. A set of network glyphs is used in the graph to facilitate the understanding and comparing of DNNs in the context of the evolution. Case studies demonstrate that DNN Genealogy provides helpful guidance in understanding, applying, and optimizing DNNs. DNN Genealogy is extensible and will continue to be updated to reflect future advances in DNNs.

Human activity recognition (HAR) using smartphones provides significant healthcare guidance for telemedicine and long-term treatment. Machine learning and deep learning (DL) techniques are widely utilized for the scientific study of the statistical models of human behaviors. However, the performance of existing HAR platforms is limited by complex physical activity. In this article, we proposed an adaptive recognition and real-time monitoring system for human activities (Ada-HAR), which is expected to identify more human motions in dynamic situations. The Ada-HAR framework introduces an unsupervised online learning algorithm that is independent of the number of class constraints. Furthermore, the adopted hierarchical clustering and classification algorithms label and classify 12 activities (five dynamics, six statics, and a series of transitions) autonomously. Finally, practical experiments have been performed to validate the effectiveness and robustness of the proposed algorithms. Compared with the methods mentioned in the literature, the results show that the DL-based classifier obtains a higher recognition rate ( $\text{95.15}\%$ , waist, and $\text{92.20}\%$ , pocket). The decision-tree-based classifier is the fastest method for modal evolution. Finally, the Ada-HAR system can monitor human activity in real time, regardless of the direction of the smartphone.

How to enable robotic compliant manipulation has become a critical problem in the robotics field. Inspired by a biomimetic adaptive control strategy, this work presents a novel representation model named human-like compliant movement primitives (Hl-CMPs) which could allow a robot to learn human-like compliant behaviours. The state-of-the-art approaches can hardly learn complete compliant profiles for a specific task. Comparatively, our model can encode task-specific parametric movement trajectories, correspondingly associated with dynamic trajectories including both impedance and feedforward force profiles. The compliant profiles are learned based on a biomimetic control strategy derived from the human motor learning in the muscle space, enabling the robot to simultaneously learn the impedance and the force while executing the movement trajectories obtained from human demonstration. Furthermore, both the kinematic and the dynamic profiles are learned in the parametric space, thus enabling the representation of a skill using corresponding parameters (i.e, task-specific parameters).

Limited evidence supports the use of early endpoints to evaluate the success of initial treatment of extranodal NK/T-cell lymphoma (ENKTCL) in the modern era. We aim to analyze progression-free survival at 24 months (PFS24) and subsequent overall survival (OS) in a large-scale multicenter cohort of patients. 1790 patients were included from the China Lymphoma Collaborative Group (CLCG) database. Subsequent OS was defined from the time of PFS24 or progression within 24 months to death. OS was compared with age- and sex-matched general Chinese population using expected survival and standardized mortality ratio (SMR). Patients who did not achieve PFS24 had a median OS of 5.3 months after progression, with 5-year OS rate of 19.2% and the SMR of 71.4 (95% CI, 62.9–81.1). In contrast, 74% patients achieved PFS24, and the SMR after achieving PFS24 was 1.77 (95% CI, 1.34–2.34). The observed OS rate after PFS24 versus expected OS rate at 5 years was 92.2% versus 94.3%. Similarly, superior outcomes following PFS24 were observed in early-stage patients (5-year OS rate, 92.9%). Patients achieving PFS24 had excellent outcome, whereas patients exhibiting earlier progression had a poor survival. These marked differences suggest that PFS24 may be used for study design and risk stratification in ENKTCL.

This paper proposes a neural approximation based model predictive control approach for tracking control of a nonholonomic wheel-legged robot in complex environments, which features mechanical model uncertainty and unknown disturbances. In order to guarantee the tracking performance of wheel-legged robots in an uncertain environment, effective approaches for reliable tracking control should be investigated with the consideration of the disturbances, including internal-robot friction and external physical interactions in the robot’s dynamical system. In this paper, a radial basis function neural network (RBFNN) approximation based model predictive controller (NMPC) is designed and employed to improve the tracking performance for nonholonomic wheel-legged robots. Some demonstrations using a BIT-NAZA robot are performed to illustrate the performance of the proposed hybrid control strategy. The results indicate that the proposed methodology can achieve promising tracking performance in terms of accuracy and stability.

Deep convolutional neural networks (CNNs) have demonstrated dominant performance in person re-identification (Re-ID). Existing CNN based methods utilize global average pooling (GAP) to aggregate intermediate convolutional features for Re-ID. However, this strategy only considers the first-order statistics of local features and treats local features at different locations equally important, leading to sub-optimal feature representation. To deal with these issues, we propose a novel weighted bilinear coding (WBC) framework for local feature aggregation in CNN networks to pursue more representative and discriminative feature representations, which can adapt to other advanced methods and improve their performance. In specific, bilinear coding is used to encode the channel-wise feature correlations to capture richer feature interactions. Meanwhile, a weighting scheme is applied on the bilinear coding to adaptively adjust the weights of local features at different locations based on their importance in recognition, further improving the discriminability of feature aggregation. To handle the spatial misalignment issue, we use a salient part net to derive salient body parts, and apply the WBC model on each part. The final representation, formed by concatenating the WBC encoded features of each part, is both discriminative and resistant to spatial misalignment. Experiments on three benchmarks including Market-1501, DukeMTMC-reID and CUHK03 evidence the favorable performance of our method against other outstanding methods.

Background and aims Repetitive transcranial magnetic stimulation (rTMS) of the left dorsolateral prefrontal cortex (DLPFC) can effectively reduce cravings in methamphetamine use disorder (MUD). However, a considerable group still fails to respond. Cognitive and emotional disturbance, as well as impulsive features, are widespread in patients with MUD and might mediate the treatment response of rTMS. The purpose of this study is to figure out whether these variables can help predicting patients' responses to rTMS treatment. Methods Ninety-seven patients with severe MUD and thirty-one gender- and age-matched healthy subjects were included. Patients were randomized to receive 20 sessions of real or sham rTMS. Intermittent theta burst protocols (iTBS) or sham iTBS were applied every weekday over the DLPFC for 20 daily sessions. Both groups received regular treatment. Craving induced by drug-related cue was measured before and after stimulation. Cognition was evaluated by using the CogState Battery. Baseline characteristics were collected through the Addiction Severity Index, Patient Health Questionnaire-9, General Anxiety Disorder Scale-7, and Barrett Impulsivity Scale-11. Results Results showed that patients with MUD have worse spatial working memory, problem-solving ability, as well as depression and anxiety symptoms compared with healthy controls. Cognition and emotion differed between responders (craving decrease ≥60%) and non-responders in real rTMS group but not in the sham group. Better cognitive and emotional functions means that patients have higher possibility for better response to real rTMS treatment. Conclusions This study suggests that cognitive, emotional and impulsive features could be used to predict the prospective treatment responses of rTMS in patients with MUD.

By highlighting important features that contribute to model prediction, visual saliency is used as a natural form to interpret the working mechanism of deep neural networks. Numerous methods have been proposed to achieve better saliency results. However, we find that previous visual saliency methods are not reliable enough to provide meaningful interpretation through a simple sanity check: saliency methods are required to explain the output of non-maximum prediction classes, which are usually not ground-truth classes. For example, let the methods interpret an image of “dog” given a wrong class label “fish” as the query. This procedure can test whether these methods reliably interpret model's predictions based on existing features that appear in the data. Our experiments show that previous methods failed to pass the test by generating similar saliency maps or scattered patterns. This false saliency response can be dangerous in certain scenarios, such as medical diagnosis. We find that these failure cases are mainly due to the attribution vanishing and adversarial noise within these methods. In order to learn reliable visual saliency, we propose a simple method that requires the output of the model to be close to the original output while learning an explanatory saliency mask. To enhance the smoothness of the optimized saliency masks, we then propose a simple Hierarchical Attribution Fusion (HAF) technique. In order to fully evaluate the reliability of visual saliency methods, we propose a new task Disturbed Weakly Supervised Object Localization (D-WSOL) to measure whether these methods can correctly attribute the model's output to existing features. Experiments show that previous methods fail to meet this standard, and our approach helps to improve the reliability by suppressing false saliency responses. After observing a significant layout difference in saliency masks between real and adversarial samples. we propose to train a simple CNN on these learned hierarchical attribution masks to distinguish adversarial samples. Experiments show that our method can improve detection performance over other approaches significantly.

The molecular cause(s) for early onset heart failure in people living with HIV-1 infection (PLWH) remains poorly defined. Herein, longitudinal echocardiography was used to assess whether NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ mice reconstituted with human hematopoietic stem cells (Hu-NSG mice) and infected with HIV-1ADA can recapitulate the salient features of this progressive human disease. Four weeks post infection, Hu-NSG mice of both sexes developed left ventricular (LV) diastolic dysfunction (DD), with 25% exhibiting grade III/IV restrictive DD with mitral regurgitation. Increases in global longitudinal and circumferential strains and declines in LV ejection fraction and fractional shortening were observed eight weeks post infection. After twelve weeks of infection, 33% of Hu-NSG mice exhibited LV dyskinesia and dyssynchrony. Histopathological analyses of hearts seventeen weeks post infection revealed coronary microvascular leakage, fibrosis and immune cell infiltration into the myocardium. These data show for the first time that HIV-1ADA-infected Hu-NSG mice can recapitulate key left ventricular cardiac deficits and pathophysiological changes reported in humans with progressive HIV-1 infection. The results also suggest that HIV-1 infected Hu-NSG mice may be a useful model to screen for pharmacological agents to blunt LV dysfunction and associated pathophysiologic causes reported in PLWH.

Dynamic network pruning achieves runtime acceleration by dynamically determining the inference paths based on different inputs. However, previous methods directly generate continuous decision values for each weight channel, which cannot reflect a clear and interpretable pruning process. In this paper, we propose to explicitly model the discrete weight channel selections, which encourages more diverse weights utilization, and achieves more sparse runtime inference paths. Meanwhile, with the help of interpretable layerwise channel selections in the dynamic network, we can visualize the network decision paths explicitly for model interpretability. We observe that there are clear differences in the layerwise decisions between normal and adversarial examples. Therefore, we propose a novel adversarial example detection algorithm by discriminating the runtime decision features. Experiments show that our dynamic network achieves higher prediction accuracy under the similar computing budgets on CIFAR10 and ImageNet datasets compared to traditional static pruning methods and other dynamic pruning approaches. The proposed adversarial detection algorithm can significantly improve the state-of-the-art detection rate across multiple attacks, which provides an opportunity to build an interpretable and robust model.

For bilateral teleoperation, the haptic feedback demands the availability of accurate force information transmitted from the remote site. Nevertheless, due to the limitation of the size, the force sensor is usually attached outside of the patient's abdominal cavity for the surgical operation. Hence, it measures not only the interaction forces on the surgical tip but also the surgical tool dynamics. In this letter, a model-free based deep convolutional neural network (DCNN) structure is proposed for the tool dynamics identification, which features fast computation and noise robustness. After the tool dynamics identification using DCNN, the calibration is performed, and the bilateral teleoperation is demonstrated to verify the proposed method. The comparison results prove that the proposed DCNN model promises prominent performance than other methods. Low computational time (0.0031 seconds) is ensured by the rectified linear unit (ReLU) function, and the DCNN approach provides superior accuracy for predicting the noised dynamics force and enable its feasibility for bilateral teleoperation.

Derived from our original nomogram study by using the risk variables from multivariable analyses in the derivation cohort of 1383 patients with extranodal NK/T-cell lymphoma, nasal-type (ENKTCL) who were mostly treated with anthracycline-based chemotherapy, we propose an easily used nomogram-revised risk index (NRI), validated it and compared with Ann Arbor staging, the International Prognostic Index (IPI), Korean Prognostic Index (KPI), and prognostic index of natural killer lymphoma (PINK) for overall survival (OS) prediction by examining calibration, discrimination, and decision curve analysis in a validation cohort of 1582 patients primarily treated with non-anthracycline-based chemotherapy. The calibration of the NRI showed satisfactory for predicting 3- and 5-year OS in the validation cohort. The Harrell’s C-index and integrated Brier score (IBS) of the NRI for OS prediction demonstrated a better performance than that of the Ann Arbor staging system, IPI, KPI, and PINK. Decision curve analysis of the NRI also showed a superior outcome. The NRI is a promising tool for stratifying patients with ENKTCL into risk groups for designing clinical trials and for selecting appropriate individualized treatment.

Methamphetamine is the most commonly used illicit drug in China. We previously demonstrated that 10 Hz rTMS over the DLPFC reduces craving in methamphetamine users. Here, we applied intermittent theta burst stimulation (iTBS), a new form of rTMS, in a large sample at four clinical centers. 126 participants (age 31.64 ± 6.33; 106 men) with severe methamphetamine use disorder according to DSM-5 were randomized to receive either iTBS or sham over the DLPFC for four weeks (20 daily sessions, 900 pulses per day). Cue-induced craving and cognitive function were assessed before and after rTMS intervention. Relapse was followed up by urine test after discharge from the rehabilitation center. iTBS significantly reduced craving and improved cognition and sleep quality. This study suggests that rTMS may be a useful and safe treatment option for methamphetamine use disorder, warranting future large scale trials.

Flame propagation pictures of a propane/air mixture are captured by a high-speed camera in a visualized constant volume combustion chamber with a diameter of 150 mm, a parallel plate height of H = ...

In recent years, deep neural networks have achieved excellent performance in many fields of artificial intelligence. The requirements for the interpretability and robustness of neural networks are also increasing. In this paper, we propose to understand the functional mechanism of neural networks by extracting critical subnetworks. Specifically, we denote the critical subnetworks as a group of important channels across layers such that if they were suppressed to zeros, the final test performance would deteriorate severely. This novel perspective can not only reveal the layerwise semantic behavior within the model but also present more accurate visual explanations appearing in the data through attribution methods. Moreover, we propose two adversarial example detection methods based on the properties of sample-specific and class-specific subnetworks, which provides the possibility for increasing the model robustness.

Electroencephalography (EEG) is a common and significant tool for aiding in the diagnosis of epilepsy and studying the human brain electrical activity. Previously, the traditional machine learning (ML)-based classifier are used to identify the seizure by extracting features from the EEG signals manually. Although the effectiveness of these contributions have already been proved, they cannot achieve multiple class classification with automatic feature extraction. Meanwhile, the identifiable EEG segment is too long to limit the capability of real-time epileptic seizure detection. In this paper, a novel deep convolutional long short-term memory (C-LSTM) model is proposed for detecting seizure and tumor in human brain and identifying two eyes statuses (open and close). It achieves to predict a result in every 0.006 seconds with a short detection duration (one second). By comparing with other two types deep learning approaches (DCNN and LSTM), the presented deep C-LSTM obtains the best performance for classifying these five classes. All of the obtained total accuracy are over 98.80%.

The surfaces of some IIB-VI semiconductors (ZnX, X = O, S, Se) are metallized by neighboring highly polar and atomically vertically aligned (VA) liquid crystal (LC) molecules. Owing to polar catastrophe, the charge carriers swarm in an extremely thin layer and the density can achieve 4.86 × 1028 m-3 close to the LC layer, which can be regarded as a 2-dimensional electron gas (2DEG). Using density functional theory (DFT), it was found that the dielectric functions of the modified layer become negative in the visible region. This indicates the semiconductor/LC platform is an ideal active plasmonic candidate, apart from the lossy metal constituents. Experimentally, after mediation with phase gratings written in the LC system, surface plasmon polaritons (SPPs) can be excited at the semiconductor surface and localized charges are gathered in an adjacent LC layer. With the help of the enhanced static electric field from the metallic surface, significantly more 2D diffraction orders in many rows and columns and a huge energy transfer between the laser beams and SPPs was observed, which is consistent with the metallization results and the bidirectional coupling between the SPPs and incident lights. The generalization of the II-VI semiconductors means the system has great promise for use in practical applications owing to the ultra-low loss. The novel insights regarding this combination with liquid crystals will be beneficial for real-time holographic displays and the study of tunable epsilon near zero points.

BACKGROUND: Risk stratification of N2 disease is vital for selecting candidates to receive invasive mediastinal staging modalities. In this study, we aimed to stratify the risk of N2 metastasis in clinical stage I lung adenocarcinoma using radiomics analysis. METHODS: Two datasets of patients with clinical stage I lung adenocarcinoma who underwent lung resection were included (training dataset, 880; validation dataset, 322). Using PyRadiomics, 1,078 computed tomography (CT)-based radiomics features were extracted after semi-automated lung nodule segmentation. In order to predict N2 status, a radiomics signature was constructed after selecting the optimal radiomics feature subset by sequentially applying minimum-redundancy-maximum-relevance and least absolute shrinkage and selection operator (LASSO) techniques. Its performance was validated in the validation dataset. RESULTS: The incidences of N2 metastasis were 8.4% and 7.1% in the training and validation datasets, respectively. Unsupervised cluster analysis revealed that radiomics features significantly correlated with lymph node status and pathological subtypes. For N2 disease prediction, five radiomics features were selected to establish the radiomics signature, which showed a significantly better predictive performance than clinical factors (P

This study aimed to explore the level of and changes in the 1,5-anhydroglucitol × glycated hemoglobin A1c/100 (AH index, AHI) associated with different glucose metabolism statuses and to evaluate the islet function and insulin sensitivity of patients with type 2 diabetes (T2DM) with different AHI levels. Of the 3562 subjects enrolled in this study, 1697 had T2DM. The disposition index (DI) was the product of islet secretion function and insulin sensitivity-related indexes. The mean AHI level was 1.0 (0.7–1.3) in the general population, while the mean AHI level in the T2DM group was 0.8 (0.5–1.2), which was significantly lower than that in the impaired glucose regulation and normal glucose tolerance group (both 1.2 (0.9–1.5), both P

The abnormal interactions of three key large-scale brain networks (default mode [DMN], salience and executive control [ECN]) were showed underlie dysfunctions in heroin addiction. Repetitive transcranial magnetic stimulation (rTMS) targeting the left dorsolateral prefrontal cortex (DLPFC) is a potential treatment for heroin addiction. It is unclear whether impaired coupling among the large-scale brain networks would be improved by rTMS in treated heroin-dependent individuals. Thirty-five heroin-dependent individuals were included in this sham-controlled, randomized study. The patients received either active or sham rTMS for 1 week. The craving for heroin and resting-state functional magnetic resonance imaging data were collected before and after 1-week rTMS. Twenty-two healthy subjects were included as controls not receiving rTMS. After 1-week rTMS, only the active rTMS group showed a significant decrease in spontaneous and heroin cue-induced craving. The coupling between left DLPFC (a key node of left ECN) and left parahippocampal gyrus (PHG, included in DMN) significantly increased for the active group with a tendency towards that of controls. The coupling between the right precentral gyrus and three key regions included in DMN (posterior cingulate cortex/precuneus and bilateral inferior parietal cortex) significantly decreased for the active group with a tendency towards that of healthy controls. For the active rTMS individuals, the left DLPFC-PHG coupling negatively correlated with the spontaneous craving and the drug cue-induced craving. It suggested that the rTMS could reduce heroin craving, which might be related to the modulation of ECN-DMN coupling. This finding might shed light on the mechanism of rTMS for heroin addicti1on treatment.

Zero-knowledge succinct arguments of knowledge (zkSNARKs) enable efficient privacy-preserving proofs of membership for general NP languages. Our focus in this work is on post-quantum zkSNARKs, with a focus on minimizing proof size. Currently, there is a 1000x gap in the proof size between the best pre-quantum constructions and the best post-quantum ones. Here, we develop and implement new lattice-based zkSNARKs in the designated-verifier preprocessing model. With our construction, after an initial preprocessing step, a proof for an NP relation of size 2^20 is just over 16 KB. Our proofs are 10.3x shorter than previous post-quantum zkSNARKs for general NP languages. Compared to previous lattice-based zkSNARKs (also in the designated-verifier preprocessing model), we obtain a 42x reduction in proof size and a 60x reduction in the prover's running time, all while achieving a much higher level of soundness. Compared to the shortest pre-quantum zkSNARKs by Groth (Eurocrypt 2016), the proof size in our lattice-based construction is 131x longer, but both the prover and the verifier are faster (by 1.2x and 2.8x, respectively). Our construction follows the general blueprint of Bitansky et al. (TCC 2013) and Boneh et al. (Eurocrypt 2017) of combining a linear probabilistically checkable proof (linear PCP) together with a linear-only vector encryption scheme. We develop a concretely-efficient lattice-based instantiation of this compiler by considering quadratic extension fields of moderate characteristic and using linear-only vector encryption over rank-2 module lattices.

Chidamide has demonstrated significant clinical benefits for patients with relapsed/refractory (R/R) PTCL in previous studies. This multi-center observational study was aimed to evaluate the objective response rate (ORR), overall survival (OS), and safety of chidamide. From February 2015 to December 2017, 548 patients with R/R PTCL from 186 research centers in China were included in the study. Among the 261 patients treated with chidamide monotherapy, ORR was 58.6% and 55 patients (21.1%) achieved complete response (CR). Among the 287 patients receiving chidamide-containing combination therapies, ORR was 73.2% and 73 patients (25.4%) achieved CR. The median OS of all patients was 15.1 months. The median OS of patients receiving chidamide monotherapy and combination therapies was 433 and 463 days, respectively. These results demonstrate a significant survival advantage of chidamide treatments as compared with international historical records. Common adverse effects (AEs) were hematological toxicities. Most AEs in both monotherapy and combined treatments were grade 1–2. No unanticipated AEs occurred. In conclusion, chidamide-based therapy led to a favorable efficacy and survival benefit for R/R PTCL. Future studies should explore the potential advantage of chidamide treatment combined with chemotherapy.

BackgroundAtelectasis is a major cause of hypoxemia during general anesthesia and postoperative pulmonary complications (PPCs).Some previous reported that the combined use of lung recruitment procedures (LRMs) and positive end-expiratory pressure (PEEP) in mechanical ventilation mode contributes to the avoidance of PPCs in patients after general anesthesia, while others suggest that the use of LRMs makes patients more susceptible to hemodynamic disturbances and lung injury, and is of limited potential to decrease the incidence of PPCs. From this perspective, controversy exists as to whether LRMs should be routinely applied to surgical patients. More importantly, corresponding clinical studies are also lacking. Therefore, this trial was conducted with the aim of solving the above problem.MethodsIn current clinical trial, patients undergoing laparoscopic gynecologic surgery with healthy lungs were randomized to the recruitment maneuvers group (RM group; 6 cm H2O PEEP and RMs) and the control group (C group; 6 cm H2O PEEP and no RMs). Lung ultrasound was performed on patients at five separate time points. During mechanical ventilation, patients in the RM group received ultrasound-guided pulmonary resuscitation when atelectasis was detected, while the C group did not intervene. Lung ultrasound scores were used to evaluate the incidence and severity of atelectasis.ResultsAfter LRMs, the incidence of atelectasis was significantly lower in the RM group (40%) than in the C group (80%) 15 minutes after arrival in the post-anesthesia care unit (PACU), and this difference did not persist for 24 hours after surgery. Meanwhile, postoperative pulmonary complications showed no difference between the two groups.ConclusionsThe combination of LRMs and PEEP decreased the incidence of atelectasis 15 minutes after admission to the PACU, but did not improve PPCs in adults with healthy lungs. Hence, for lung-healthy patients undergoing gynecological laparoscopic surgery, we do not recommend routine recruitment maneuvers. Trial registration (prospectively registered)ChiCTR2000033529. Registered on June 4, 2020.

Water vapor is one of the most important greenhouse gases in the world. There are many techniques that can measure water vapor directly or remotely. In this work, we first study the Global Positioning System (GPS)- and the Global Navigation Satellite System (GLONASS)-derived Zenith Wet Delay (ZWD) time series based on 11 years of the second reprocessing campaign of International Global Navigation Satellite Systems (GNSS) Service (IGS) using 320 globally distributed stations. The amount of measurement, the local environment, and the antenna radome are shown to be the main factors that affect the GNSS ZWDs and the corresponding a posteriori formal errors. Furthermore, antenna radome is able to effectively reduce the systematic bias of ZWDs and a posteriori formal errors between the GPS- and GLONASS-based solutions. With the development of the GLONASS, the ZWD differences between the GPS- and the GLONASS-based solutions have gradually decreased to sub-mm-level after GLONASS was fully operated. As the GPS-based Precipitable Water Vapor (PWV) is usually used as the reference to evaluate the other PWV products, the PWV consistency among several common techniques is evaluated, including GNSSs, spaceborne sensors, and numerical products from the European Center for Medium-Range Weather Forecasts (ECMWF). As an example of the results from a detailed comparison analysis, the long-term global analysis shows that the PWV obtained from the GNSS and the ECMWF have great intra-agreements. Based on the global distribution of the magnitude of the PWV and the PWV drift, most of the techniques showed superior agreement and proved their ability to do climate research. With a detailed study performed for the ZWDs and PWV on a long-term global scale, this contribution provides a useful supplement for future research on the GNSS ZWD and PWV.

Ship detection in synthetic aperture radar (SAR) images plays an important role in remote sensing, but it is still full of challenges in the deep learning area. The primary problem is that ships in SAR images have different sizes and orientations. The off-the-shelf detectors are not able to adapt to the situation. Recurrent feature pyramid networks are presented to detect ships with different sizes especially the small ones. Rotatable region proposal network is used for locating ships with a tighter rectangle. Rotatable anchors with sizes, aspect ratios, and angles are designed according to the distribution of ships in dataset. Multiratio region-of-interest pooling is used for projecting arbitrary-oriented proposals to fixed length vectors. Angle-related intersection-of-unit (ArIoU) is used for evaluating the intersection of rotatable proposals. ArIoU can be an indicator for nonmaximum suppression (NMS) and also is used for preparing negative and positive proposals. A loss function is proposed to compute loss between bounding boxes. The sinusoidal function is used for solving the problem of unstable angle. We also use a dataset called SSDD+ (SAR ship detection dataset plus) to evaluate different methods. Experiments based on SSDD+ show that our method achieves state-of-the-art performance. The dataset and the code will be public at https://zhuanlan.zhihu.com/p/143794468.

This paper combines the advantages of regular segmentation and statistical segmentation, proposing an incomplete maximum matching segmentation method based on semantics. On the basis of ensuring time consumption, the new method solves the defect of word adhesion in the maximum matching algorithm. The innovative work includes: in the preprocessing stage, the forward semantic similarity dictionary is constructed to realize the follow-up word recognition. In the stage of word segmentation, the formula of three feature weight is proposed to redefine the segmentation principle. The experimental results show that the new method has a certain improvement in the accuracy and recall rate, which is suitable for the field of text processing.

Aquaculture has undergone rapid development in the past decades. It provides a large part of high-quality protein food for humans, and thus, a sustainable aquaculture industry is of great importance for the worldwide food supply and economy. Along with the quick expansion of aquaculture, the high fish densities employed in fish farming increase the risks of outbreaks of a variety of aquatic diseases. Such diseases not only cause huge economic losses, but also lead to ecological hazards in terms of pathogen spread to marine ecosystems causing infection of wild fish and polluting the environment. Thus, fish health is essential for the aquaculture industry to be environmentally sustainable and a prerequisite for intensive aquaculture production globally. The wide use of antibiotics and drug residues has caused intensive pollution along with risks for food safety and increasing antimicrobial resistance. Vaccination is the most effective and environmentally friendly approach to battle infectious diseases in aquaculture with minimal ecological impact and is applicable to most species of farmed fish. However, there are only 34 fish vaccines commercially available globally to date, showing the urgent need for further development of fish vaccines to manage fish health and ensure food safety. Plant genetic engineering has been utilized to produce genetically modified crops with desirable characteristics and has also been used for vaccine production, with several advantages including cost-effectiveness, safety when compared with live virus vaccines, and plants being capable of carrying out posttranslational modifications that are similar to naturally occurring systems. So far, plant-derived vaccines, antibodies, and therapeutic proteins have been produced for human and animal health. However, the development of plant-made vaccines for animals, especially fish, is still lagging behind the development of human vaccines. The present review summarizes the development of fish vaccines currently utilized and the suitability of the plant-production platform for fish vaccine and then addresses considerations regarding fish vaccine production in plants. Developing fish vaccines by way of plant biotechnology are significant for the aquaculture industry, fish health management, food safety, and human health.

BackgroundThis study aimed to explore the level and changes in handgrip strength among preclinical Alzheimer’s disease (AD) and AD patients and to evaluate the association between handgrip strength and cognitive function.MethodsA total of 1431 participants from the memory clinic of Shanghai JiaoTong University Affiliated Sixth People’s Hospital and community were enrolled in the final analysis, including 596 AD, 288 mild cognitive impairment (MCI), and 547 normal individuals (NC). All participants received a comprehensive neuropsychological assessment. Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment-Basic (MoCA-BC), and the Chinese version of Addenbrooke’s Cognitive Examination III (ACE-III-CV) were used as cognitive tests. The receiver operating characteristic curve (ROC) was plotted to assess the power of handgrip strength as a screening measure to discriminate AD and MCI.ResultsThe results showed that handgrip strength in the MCI group was significantly lower than that of NC group, and the AD group had a further decline (bothP ). Multivariate logistic regression was performed with the handgrip strength quartiles, and the results showed that the ORs of AD for increasing levels of handgrip strength were 1.00, 0.58 (0.46–0.78), 0.51 (0.36–0.73), and 0.50 (0.35–0.68), showing a decreasing trend (Pfor trend<0.01). The ROC curve demonstrated that the handgrip strength cutoff points for the identification of AD were 16.8 and 20.7 kg among the female participants above and under 70 yrs and 24.4 and 33.3 kg for the male participants above and under 70 yrs, respectively. Similarly, for the identification of MCI, cutoff points were 17.5 and 21.9 kg for females above 70 yrs and under 70 yrs, and 25.8 and 36.2 kg for males above 70 yrs and under 70 yrs, respectively.ConclusionsOur study provided the further knowledge on the relationship between noncognitive features and cognition in populations with differing cognitive status, revealed that the stronger handgrip strength was associated with better performances on cognitive function. It can be speculated that handgrip strength can help early recognition of Chinese AD patients.

BackgroundPrevious studies have identified the treatment effect of repetitive transcranial magnetic stimulation (rTMS) on cravings of patients with methamphetamine use disorder (MUD). However, the mechanism underlying the treatment effect remains largely unknown. A potential candidate mechanism could be that rTMS over the dorsolateral prefrontal cortex (DLPFC) modulates the attention bias to methamphetamine-related cues. The purpose of this study is therefore to determine the modulation of rTMS on methamphetamine-related attention bias and the corresponding electrophysiological changes.MethodsForty-nine patients with severe MUD were included for analysis. The subjects were randomized to receive the active intermittent theta-burst stimulation (iTBS) or sham iTBS targeting DLPFC for 20 sessions. Participants performed the Addiction Stroop Task before and after the treatment while being recorded by a 64-channel electroencephalogram. Baseline characteristics were collected through the Addiction Severity Index.ResultsPost-treatment evaluations showed a reduced error rate in discriminating the color of methamphetamine words in the active iTBS group compared with the sham iTBS group. Following rTMS treatment, we found the significant time-by-group effect for the N1 amplitude (methamphetamine words > neutral words) and P3 latency (methamphetamine words > neutral words). The change of N1 amplitude was positively correlated with cravings in the active group. Moreover, reduced power of neural oscillation in the beta band, manifesting at frontal central areas, was also found in the active group.ConclusionThis study suggests that attention bias and the beta oscillation during the attentional processing of methamphetamine words in patients with MUD could be modulated by iTBS applied to left DLPFC.