Search

Purpose Low back pain (LBP) is a long-lasting and chronic symptom without any exact cause. This study attempts to propose a new staging system based on the original grading system combined with pathological results and clinical symptoms to better clarify the dynamic evolution of LBP related to cartilage degeneration during facet joint osteoarthritis (FJOA). To explore a potential target for diagnosis, treatment, and drug intervention of facet joint osteoarthritis related LBP via protecting chondrocytes. Materials and methods All the facet joints were divided into 4 groups according to our new degenerative staging system based on Weishaupt grade, CT and MRI. Collect the facet joint samples from patients whom suffered lumbar fusion surgery for lumbar disc herniation. Molecular biology experiments were used to explore the effect of Wnt16 on the degeneration of facet joints. Micro-CT examination and pain stimulation test checked the biological function of Wnt16 in rats. Results Wnt16 was significantly increased and more aggregated in the facet joint chondrocytes in the Phase III and Phase IV, which is consistent with the pathological findings of cartilage degeneration (OARSI). We found that Wnt16 participated in the regulation of FJOA via Wnt/β-catenin pathway in vitro, which was inhibited by specific inhibitor DKK1. The rats, rich expressed Wnt16, showed higher paw withdrawal thresholds and prolonged paw withdrawal latency to FJOA related LBP. Micro-CT examination for the lumbar spine of rats showed Wnt16 protected the chondrocytes from FJOA. Conclusions This study defined a new staging system for LBP related cartilage degeneration of facet joint based on the original grading system combined with pathological results and clinical symptoms. Wnt16 is expected to be a potential target for treatment of FJOA via protecting chondrocytes.

The valley degrees of freedom of Bloch electrons provide a proper platform to realize information storage and processing. Using first principles calculations, we propose that the FeClBr monolayer is a ferromagnetic semiconductor with spontaneous valley polarization owing to the combined effect of magnetic exchange interaction and spin–orbit coupling effect. The FeClBr monolayer shows perpendicular magnetic anisotropy, a high Curie temperature of 651 K and a large valley splitting of 188 meV, which are beneficial for the practical applications in valleytronics. Then, the anomalous valley Hall effect can be realized under an in-plane electrical field due to the valley-contrasting berry curvature. According to the optical selectivity rule, the different valleys at K and K− points in momentum space can be excited by the circularly polarized light in honeycomb structures; however, the FeClBr monolayer can also respond to the linear light. Therefore, the valley degree of freedom of the FeClBr monolayer can be modulated by circularly polarized light, linear light and hole doping. Our work enriches the library of valley materials and provides a candidate for the study of spintronics and valleytronics field.

Two-dimensional (2D) ferromagnetic materials with high spin polarization are highly desirable for spintronic devices. 2D Janus materials exhibit novel properties due to their broken symmetry. However, the electronic structure and magnetic properties of 2D Janus magnetic materials with high spin polarization are still unclear. Inspired by the successful synthesis of a ferromagnetic FeCl2 monolayer and 2D Janus MoSSe and WSSe, we systematically study the electronic structure and magnetic properties of Janus FeXY (X, Y = Cl, Br, and I, X ≠ Y) monolayers. Based on the Goodenough-Kanamori-Anderson theory, the ferromagnetism stems from the superexchange interaction mediated by Fe-X/Y-Fe bonds. The band gaps of spin-up channels are large enough (>4 eV) to prevent spin flipping, which is beneficial for spintronic devices. Additionally, the sizable magnetocrystalline anisotropy energy (MAE) indicates that Janus FeXY monolayers are suitable for information storage. More importantly, the half-metallic character is still kept in Janus FeXY monolayers, and their magnetic properties are enhanced by the biaxial compressive strain. The MAE of FeClI and FeBrI increases by 1 order of magnitude, and the Curie temperature of FeXY monolayers enhances by 100%. These results provide an example of the 2D Janus half-metallic materials and enrich the 2D magnetic material library.

Background: Roxadustat, a hypoxia-inducible factor prolyl-hydroxylase inhibitor (HIF-PHI), has been used to treat anemia in patients with chronic kidney disease (CKD). However, its safety and efficacy remain controversial.Methods: The PubMed, EMBASE, Science Citation Index, Cochrane Central Register of Controlled Trials, and Clinical Trial Registries databases were searched for relevant studies published up to April 2021. We identified randomized controlled trials (RCTs) comparing roxadustat with placebo or erythropoiesis-stimulating agents (ESAs) in anemia patients with CKD with or without dialysis.Results: Eleven studies including 6,631 patients met the inclusion criteria. In non-dialysis-dependent (NDD-) and dialysis-dependent (DD-) CKD patients, the total adverse events were not significantly different between the roxadustat and control (placebo for NDD-CKD patients and ESA for DD-CKD patients) groups [relative risk (RR) = 1.02, 95% confidence interval (CI) = 1.00, 1.04, P = 0.08, and RR = 1.22, 95% CI = 0.91, 1.64, P = 0.18, respectively], and the trial sequential analysis (TSA) confirmed the result in the NDD-CKD groups. No significant differences in hyperkalemia and infection incidences were found between roxadustat and placebo in the DD-CKD groups. The pooled results showed that roxadustat significantly increased the hemoglobin response rate compared with placebo in the NDD-CKD group and had an effect similar to that of ESA in the DD-CKD group. However, iron metabolism parameters did not seem to be obviously optimized by roxadustat.Conclusion: Roxadustat can be safely used in CKD patients. Oral roxadustat was more effective than placebo as a therapy for anemia in NDD-CKD patients and non-inferior to ESA in correcting anemia in DD-CKD patients. However, additional clinical trials are still needed to further prove whether roxadustat can optimize iron metabolism.

Research on flourishing public bike-sharing systems has been widely discussed in recent years. In these studies, many existing works focus on accurately predicting individual stations in a short time. This work, therefore, aims to predict long-term bike rental/drop-off demands at given bike station locations in the expansion areas. The real-world bike stations are mainly built-in batches for expansion areas. To address the problem, we propose LDA (Long-Term Demand Advisor), a framework to estimate the long-term characteristics of newly established stations. In LDA, several engineering strategies are proposed to extract discriminative and representative features for long-term demands. Moreover, for original and newly established stations, we propose several feature extraction methods and an algorithm to model the correlations between urban dynamics and long-term demands. Our work is the first to address the long-term demand of new stations, providing the government with a tool to pre-evaluate the bike flow of new stations before deployment, this can avoid wasting resources such as personnel expense or budget. We evaluate real-world data from New York City’s bike-sharing system, and show that our LDA framework outperforms baseline approaches.

All-reduce is the key communication primitive used in distributed data-parallel training due to the high performance in the homogeneous environment. However, All-reduce is sensitive to stragglers and communication delays as deep learning has been increasingly deployed on the heterogeneous environment like cloud. In this paper, we propose and analyze a novel variant of all-reduce, called partial-reduce, which provides high heterogeneity tolerance and performance by decomposing the synchronous all-reduce primitive into parallel-asynchronous partial-reduce operations. We provide theoretical guarantees, proving that partial-reduce converges to a stationary point at the similar sub-linear rate as distributed SGD. To enforce the convergence of the partial-reduce primitive, we further propose a dynamic staleness-aware distributed averaging algorithm and implement a novel group generation mechanism to prevent possible update isolation in heterogeneous environments. We build a prototype system in the real production cluster and validate its performance under different workloads. The experiments show that it is 1.21x-2x faster than other state-of-the-art baselines.

With the ever-evolving concerns on privacy protection, vertical federated learning (FL), where participants own non-overlapping features for the same set of instances, is becoming a heated topic since it enables multiple enterprises to strengthen the machine learning models collaboratively with privacy guarantees. Nevertheless, to achieve privacy preservation, vertical FL algorithms involve complicated training routines and time-consuming cryptography operations, leading to slow training speed. This paper explores the efficiency of the gradient boosting decision tree (GBDT) algorithm under the vertical FL setting. Specifically, we introduce VF^2Boost, a novel and efficient vertical federated GBDT system. Significant solutions are developed to tackle the major bottlenecks. First, to handle the deficiency caused by frequent mutual-waiting in federated training, we propose a concurrent training protocol to reduce the idle periods. Second, to speed up the cryptography operations, we analyze the characteristics of the algorithm and propose customized operations. Empirical results show that our system can be 12.8-18.9 times faster than the existing vertical federated implementations and support much larger datasets.

The appeal of serverless (FaaS) has triggered a growing interest on how to use it in data-intensive applications such as ETL, query processing, or machine learning (ML). Several systems exist for training large-scale ML models on top of serverless infrastructures (e.g., AWS Lambda) but with inconclusive results in terms of their performance and relative advantage over "serverful" infrastructures (IaaS). In this paper we present a systematic, comparative study of distributed ML training over FaaS and IaaS. We present a design space covering design choices such as optimization algorithms and synchronization protocols, and implement a platform, LambdaML, that enables a fair comparison between FaaS and IaaS. We present experimental results using LambdaML, and further develop an analytic model to capture cost/performance tradeoffs that must be considered when opting for a serverless infrastructure. Our results indicate that ML training pays off in serverless only for models with efficient (i.e., reduced) communication and that quickly converge. In general, FaaS can be much faster but it is never significantly cheaper than IaaS.

In this work, we prepared a layer of hydroxyapatite (HA) coating on the surface of titanium alloy via hydrothermal process. FTIR, SEM, Raman, XRD and electrochemical tests were used to analyze the structure of prepared HA coating. The prepared HA coating on the surface of titanium alloy demonstrated superior corrosion resistances. Antibacterial experiments proved that HA coating could adsorb drugs and act as a carrier for antibiotics to achieve inhibition of bacteria. The experimental results indicated that HUVEC cells could adhere to the surface of substrates. Animal experiments also confirmed that HA coating has good biocompatibility and promote the formation of new bone.

Excessive alcohol consumption is positively related to osteoporosis, and its treatment strategies are poorly developed. Disulfiram inhibits receptor activator of nuclear factor kappa-B ligand (RANKL)-induced osteoclastogenesis; however, whether it can be used for ethanol-induced osteoclastogenesis and its underlying mechanism are still unclear. In this study, we demonstrated that ethanol promoted RANKL-induced osteoclast formation and bone resorption, whereas, disulfiram suppressed ethanol-induced osteoclastogenesis by abrogating the expression of nuclear factor of activated T cell c1 (NFATc1) in vitro. Further analysis revealed that aldehyde dehydrogenase 1A1 (ALDH1A1) is important for the expression of NFATc1, the master regulator of osteoclast differentiation. Furthermore, we showed that disulfiram protected ethanol-induced osteoporosis in vivo. Overall, our study provides promising evidence that disulfiram can be used as a treatment strategy for alcohol-related osteoporosis via the ALDH1A1T-NFATc1 axis.

Osteoarthritis (OA), which is characterized as a common degenerative joint disease, is presently the most prevalent chronic degenerative joint disease. Accumulating evidence has shown a biological function for Garcinol in a variety of diseases; however, whether it could be used to treat OA remains unclear. In this study, we explored the protective effects of garcinol on the progression of OA and explored the underlying mechanism. In vitro, garcinol reduced the expression of pro-inflammatory cytokines, such as IL-6 and tumor necrosis factor alpha (TNF-α). It also decreased the expression of inducible nitric oxide synthase (iNOS), as well as cyclooxygenase-2 (COX-2). Furthermore, garcinol inhibited the expression of thrombospondin motifs 5(ADAMTS5) and metalloproteinase (MMPs), both of which regulate extracellular matrix degradation. These changes could be attributed to garcinol-related suppression of the IL-1β-induced NF-κB signaling pathway. Moreover, we investigated the protective effects of garcinol on the surgical destabilization of the medial meniscus (DMM) of the mouse, an in vivo model of OA. Taken together, our data suggest garcinol as a potential future agent for the treatment of OA.

Simplified equivalent static methods are widely used in seismic design and analysis of underground structures. The relative shear deformation of the soil is usually considered as the main earthquake load in the existing simplified methods, including the free-field racking deformation method, flexible coefficient method, response displacement method, response acceleration method, and pushover analysis method. However, the inertia force of the overburden soil caused by the vertical earthquake motion does not attract sufficient attention in many design specifications. The recent studies revealed that the large vertical inertia force made a significant influence on the shear strength and deformation capacity of the support components in the underground structures, especially for the shallow buried ones. Firstly, based on the earthquake damage investigation of shallow buried rectangular underground structures and the analysis model of response displacement method (RDM), this paper proposes the vertical inertia force-response displacement method (VIF-RDM), in which the vertical inertia force of overburden soil is taken into account. The calculation methods of two critical parameters of the VIF-RDM are described in detail, including the coefficients of the foundation springs around the structure and the maximum vertical inertia force of the overburden soil. Afterwards, the integral vertical inertia force-response displacement method (IVIF-RDM) is put forward to reduce the computational complexity and calculation error of the foundation springs. The surrounding soil model is directly used to calculate the equivalent earthquake loads, which are caused by the relative displacement of the free field. The two proposed simplified methods and RDM are compared with the time-history analysis method (THAM) in a real underground structure. The results show a significant difference in the axial force of the central column between the two proposed methods and RDM. The proposed methods reflect the internal forces of the structure under the strong earthquake more accurately. Therefore, it can be used to evaluate the seismic safety performance of the shallow buried rectangular underground structures in the future.

Postmenopausal osteoporosis is the most common form of primary osteoporosis, and the incidence of the condition is rapidly increasing. In consideration of the limitations of current therapeutic opt...

Osteolytic diseases are characterized by an increase in the number and/or activity of bone-resorbing osteoclasts. Identification of natural compounds that can suppress osteoclast formation and function is crucial for the prevention and treatment of osteolytic diseases. Vitexin, a naturally-derived flavonoid extracted from various medicinal plant species, demonstrates a broad range of pharmacological properties including anticancer and anti-inflammatory effects. Here in this study, we showed that vitexin exerts antiosteoclastogenic effects by directly inhibiting receptor activator of nuclear factor-κB ligand (RANKL)-induced osteoclast formation and bone resorption in vitro and protected against lipopolysaccharide (LPS)-induced inflammatory osteolysis in vivo. Vitexin suppressed the early activation of ERK and p38 MAPK pathways in response to RANKL thereby attenuating the downstream induction of c-Fos and NFATc1, and abrogating the expression of osteoclast marker genes. Collectively, these results provide evidence for the therapeutic application of vitexin in the treatment of osteoclast-mediated bone lytic diseases.

Introducing sulfur species into Pt catalysts has been proven to be an effective method to improve their performance in various reactions. However, the role of sulfur addition on Pt catalysts catalyzing propane dehydrogenation (PDH) is still not clear. This work combines catalyst characterizations, catalytic kinetics studies and DFT calculations to understand the influence of H2S addition in the feed on Pt/θ-Al2O3 catalyzed PDH from a mechanistic perspective. With the addition of a trace amount of H2S in the feed (3 ppm), the propylene selectivity increases from 79% to 96% and the deactivation factor decreases from 33% to 18%, at the expense of a slight activity loss. The improved catalytic performance can be partially attributed to the electron transfer from sulfur species to Pt atoms as indicated by CO-DRIFTS results. DFT calculations show that H2S could be the dominating sulfur species on Pt particles and donate electrons to Pt atoms, which further proves that sulfur species improve the catalytic performance by donating electrons. In addition, the repulsion between sulfur species and C3 hydrocarbons also explains the improved catalytic performance. © 2019. This is the authors' accepted and refereed manuscript to the article. The final authenticated version is available online at: https://doi.org/10.1039/C8CY02393J

The seismic response characteristics of underground structures are very different from those of aboveground structures. Great efforts have been made in investigating the influence factors on the seismic responses of rectangular underground structures, while the theoretical analyses and comprehensive understandings remain insufficient in the current research. In order to shed light on the seismic response characteristics of the shallow buried rectangular underground structure, an extended parametric study, based on a real underground structure, the Daikai subway station in Japan was conducted in the present paper. In particular, a two-dimensional soil-structure system was adopted for the dynamic time-history analysis. The equivalent linear model was used to consider the nonlinear behaviors of the soil elements, and the elastic model was used to simulate the structure elements. A variety of critical factors that may significantly affect the seismic responses, such as: (i) earthquake load condition, (ii) structure inertia effect, (iii) soil inertia effect, (iv) soil-structure relative stiffness, and (v) soil-structure interface properties were comprehensively investigated in this study. The variations of the horizontal relative deformation between the top and bottom slabs of the structure and the internal forces of four representative sections that were most concerned in seismic design were calculated and compared in different cases, and the soil pressures and shear forces around the structure were also systematically discussed. The numerical results indicated that the seismic responses were obviously affected by the soil inertia effect and soil-structure relative stiffness, and more attentions should be paid to these critical factors during the seismic design of the underground structure. The presented results can lead to better understandings on the seismic responses of the shallow buried rectangular underground structure and can provide some improvements to the existing simplified methods and guide the seismic design of the underground structure.

Two-dimensional magnetic materials offer an ideal platform to study the fundamental research and potential applications in spintronics. However, it is still a challenging issue to raise their magnetic anisotropy energy and Curie temperature. Based on the density functional theory and the Monte Carlo method, we investigate the electron-mediated electronic structures and magnetic properties in a FeCl2/Ca2N heterostructure. It is found that a Ca2N monolayer devotes about 0.46 electrons/f.u. to a FeCl2 monolayer due to the low work function. This provides a high electron doping density of 4.28 × 1014 cm−2 and avoids the disorder effect of the traditional doping methods simultaneously such as the intercalation, adsorption, and substitution. Moreover, the magnetocrystalline anisotropy energy and Curie temperature of the FeCl2/Ca2N heterostructure enhance 66% and 100% compared with the FeCl2 monolayer, respectively. Our findings provide a practical and effective avenue to modulate the magnetic properties of two-dimensional magnetic materials.

Fibroblast-like synoviocytes (FLSs) play an essential role in rheumatoid arthritis (RA) by promoting synovitis, pannus growth and cartilage/bone destruction. Increased proliferation, migration and invasion of FLSs greatly contribute to RA initiation and progression. Dual-specificity tyrosine-regulated kinase 1A (Dyrk1A), a serine/threonine kinase, regulates MAPK pathway activation, and governs the proliferation and differentiation of neuronal progenitor cells and cancer cells. Till now, the expression and possible function of Dyrk1A in RA FLSs have not been explored. In this study, we detected an increased expression of Dyrk1A both in the synovial tissues of RA patients and in a TNF-α-induced FLSs activation model. CCK-8 and Edu assays revealed that Dyrk1A knockdown inhibited TNF-α-induced FLSs proliferation. Moreover, inhibiting Dyrk1A expression apparently prevented the migration and invasion capability of FLSs accompanied by a decreased MMP-3 and -9 expression. To investigate the molecular mechanism through which Dyrk1A modulates FLSs activities, we evaluated the effects of Dyrk1A on Spry2, a negativity modulator of ERK MAPK pathway. Western blot assay demonstrated that Dyrk1A silencing significantly increased Spry2 expression and suppressed the phosphorylation of ERK in TNF-α-treated FLSs. Taken together, our results indicated that Dyrk1A might promote FLSs proliferation, migration and invasion by suppressing Spry2 expression and activating the ERK MAPK signaling pathway in RA.

As a prevalent technique for checking the correctness of software, software verification has achieved a significant progress in the past decades, reaching a point where a large number of verification algorithms and tools are available and sophisticated enough to handle the large-scale industrial software. However, it remains a difficult task to select a suitable verification algorithm or tool for the software at hand, given the fact that the underlying algorithms are diverse and the performance tradeoffs are hard to accurately characterize. In this paper, we study the algorithm selection problem for software verification, and propose a novel algorithm selection model based on the Long Short Term Memory network (LSTM). Our solution employs word2vec to obtain the embedding representation of the code, avoiding constructing the software features manually. We also propose a novel approach to construct the adversarial code examples in order to solve the sparsity and data imbalance problem. The experimental evaluations on the latest available dataset show that our solution improves the prediction accuracy by about 7% compared with the state-of-the-art selection algorithm.

In this paper, we propose an efficient GPU-training framework for the large-scale wide models, named cuWide. To fully benefit from the memory hierarchy of GPU, cuWide applies a new flow-based schema for training, which leverages the spatial and temporal locality of wide models to drastically reduce the amount of communication with GPU global memory. Comprehensive experiments show that cuWide can be up to more than 20× faster than the state-of-the-art GPU solutions and multi-core CPU solutions.

Colonoscopy is an effective tool for early screening of colorectal diseases. However, the application of colonoscopy in distinguishing different intestinal diseases still faces great challenges of efficiency and accuracy. Here we constructed and evaluated a deep convolution neural network (CNN) model based on 117 055 images from 16 004 individuals, which achieved a high accuracy of 0.933 in the validation dataset in identifying patients with polyp, colitis, colorectal cancer (CRC) from normal. The proposed approach was further validated on multi-center real-time colonoscopy videos and images, which achieved accurate diagnostic performance on detecting colorectal diseases with high accuracy and precision to generalize across external validation datasets. The diagnostic performance of the model was further compared to the skilled endoscopists and the novices. In addition, our model has potential in diagnosis of adenomatous polyp and hyperplastic polyp with an area under the receiver operating characteristic curve of 0.975. Our proposed CNN models have potential in assisting clinicians in making clinical decisions with efficiency during application., Precision Clinical Medicine, 4 (2)

Black-box optimization (BBO) has a broad range of applications, including automatic machine learning, engineering, physics, and experimental design. However, it remains a challenge for users to apply BBO methods to their problems at hand with existing software packages, in terms of applicability, performance, and efficiency. In this paper, we build OpenBox, an open-source and general-purpose BBO service with improved usability. The modular design behind OpenBox also facilitates flexible abstraction and optimization of basic BBO components that are common in other existing systems. OpenBox is distributed, fault-tolerant, and scalable. To improve efficiency, OpenBox further utilizes "algorithm agnostic" parallelization and transfer learning. Our experimental results demonstrate the effectiveness and efficiency of OpenBox compared to existing systems.

The ensemble of deep neural networks has been shown, both theoretically and empirically, to improve generalization accuracy on the unseen test set. However, the high training cost hinders its efficiency since we need a sufficient number of base models and each one in the ensemble has to be separately trained. Lots of methods are proposed to tackle this problem, and most of them are based on the feature that a pre-trained network can transfer its knowledge to the next base model and then accelerate the training process. However, these methods suffer a severe problem that all of them transfer knowledge without selection and thus lead to low diversity. As the effect of ensemble learning is more pronounced if ensemble members are accurate and diverse, we propose a method named Efficient Diversity-Driven Ensemble (EDDE) to address both the diversity and the efficiency of an ensemble. To accelerate the training process, we propose a novel knowledge transfer method which can selectively transfer the previous generic knowledge. To enhance diversity, we first propose a new diversity measure, then use it to define a diversity-driven loss function for optimization. At last, we adopt a Boosting-based framework to combine the above operations, such a method can also further improve diversity. We evaluate EDDE on Computer Vision (CV) and Natural Language Processing (NLP) tasks. Compared with other well-known ensemble methods, EDDE can get highest ensemble accuracy with the lowest training cost, which means it is efficient in the ensemble of neural networks., Comment: 12 pages, 8 figures

Hyperparameter optimization (HPO) is a fundamental problem in automatic machine learning (AutoML). However, due to the expensive evaluation cost of models (e.g., training deep learning models or training models on large datasets), vanilla Bayesian optimization (BO) is typically computationally infeasible. To alleviate this issue, Hyperband (HB) utilizes the early stopping mechanism to speed up configuration evaluations by terminating those badly-performing configurations in advance. This leads to two kinds of quality measurements: (1) many low-fidelity measurements for configurations that get early-stopped, and (2) few high-fidelity measurements for configurations that are evaluated without being early stopped. The state-of-the-art HB-style method, BOHB, aims to combine the benefits of both BO and HB. Instead of sampling configurations randomly in HB, BOHB samples configurations based on a BO surrogate model, which is constructed with the high-fidelity measurements only. However, the scarcity of high-fidelity measurements greatly hampers the efficiency of BO to guide the configuration search. In this paper, we present MFES-HB, an efficient Hyperband method that is capable of utilizing both the high-fidelity and low-fidelity measurements to accelerate the convergence of HPO tasks. Designing MFES-HB is not trivial as the low-fidelity measurements can be biased yet informative to guide the configuration search. Thus we propose to build a Multi-Fidelity Ensemble Surrogate (MFES) based on the generalized Product of Experts framework, which can integrate useful information from multi-fidelity measurements effectively. The empirical studies on the real-world AutoML tasks demonstrate that MFES-HB can achieve 3.3-8.9x speedups over the state-of-the-art approach --- BOHB.

Bioceramics have been used in orthopedic surgery for several years. Magnesium (Mg) is an essential element in bone tissue and plays an important role in bone metabolism. Mg-doped bioceramics has attracted the attention of researchers recently. However, the optimal doping amount of Mg in β-TCP and the immunomodulatory property of Mg-doped β-TCP (Mg-TCP) have not been determined yet. In this study, β-TCP scaffolds doped with different contents of magnesium oxide (0 wt%, 1 wt%, 3 wt%, and 5 wt%) with gyroid structure were printed by digital light processing (DLP) method, and the physicochemical and biological functions were then investigated. Mg-doping improved the physicochemical properties of the β-TCP scaffolds. In vitro experiments confirmed that the doping of Mg in β-TCP scaffolds promoted the osteogenic differentiation of bone marrow mesenchymal stem cells (BMSCs) and angiogenic differentiation of endothelial progenitor cells (EPCs), where the 5Mg-TCP has the optimal properties when using the "one cell type" method. It was also found that all Mg-TCP facilitated the polarization of RAW264.7 cells to the M2 phenotype, especially the 3Mg-TCP. However, 3Mg-TCP displayed the optimal osteogenic and angiogenic potential when using a "multiple cell type" method, which referred to culturing the BMSCs or EPCs in the macrophage-conditioned medium. Finally, the in vivo experiments were conducted and the results confirmed that the 3Mg-TCP scaffolds possessed the satisfying bone defect repair capability both after 6 and 12 weeks of implantation. This study suggests that 3Mg-TCP scaffolds provide the optimal biological performance and thus have the potential for clinical translation.

Aerodynamic drag reduction is a key element for the design of aircrafts, and it is also considered to be affected by the flow velocity. Herein, the influence of high flow velocity on the drag reduction induced by the surface microstructure inspired by a cross-section of barchan dune was investigated by the computational fluid dynamics method in this work. Overall, the drag reduction ratio was decreased while the pressure drag and viscous resistance enhanced simultaneously with the augmentation of flow velocity. Otherwise, drag analysis revealed that the total drag was a power function of flow velocity, which meant that the effect of flow velocity on drag was extremely fierce. Additionally, the microstructure improved the thickness of the boundary layer with a growth rate of 14.2%, and then reduced the viscosity resistance with limits during the development process of flow velocity. Furthermore, the micro-vortex caused by the surface microstructure provided the reverse wall shear stress, with the maximum value ranging from −4.77 Pa to −51.27 Pa, and then reduced the velocity gradient above the microstructure, thereby improving the drag reduction. However, both Reynolds-averaged Navier-Stokes (RANS) and large eddy simulation (LES) calculations showed that the excessive velocity could lead to the dissipation of micro-vortex, which augmented the contact area between the fluid and the surface, resulting in the enlargement of viscous resistance. Finally, it was confirmed that the variation of surface microstructure height had a significant influence on drag reduction at high flow velocity. The underlying mechanism of drag reduction could also provide theoretical guidance for the design and optimization of drag reduction coatings in aeronautical applications.

The discovery of two-dimensional magnetic and ferroelectric materials provides a feasible opportunity to realize the nonvolatile electrical control of magnetic properties in van der Waals multiferroic heterostructures. Here, based on the density of functional theory, we study the electronic structures and magnetic properties of the Janus FeBrI/In2S3 multiferroic heterostructure. The Curie temperature and magnetic anisotropy energy (MAE) can be modulated by the ferroelectric polarization due to the interfacial charge redistribution. According to the orbital-resolved magnetocrystalline anisotropy energy ( EMCA) and the second-order perturbation theory, the difference of EMCA mainly stems from the change in electronic states near the Fermi level, which is associated with interfacial charge redistribution in different ferroelectric polarization directions. These findings not only reveal the physical mechanism of the change of MAE in detail but also suggest a feasible strategy to modulate the magnetic properties.

Photothermocatalytic CO2 reduction as the channel of the energy and environmental issues resolution has captured persistent attention in recent years. In2O3 has been prompted to be a potential photothermal catalyst in this sector on account of unique physicochemical properties. However, different from the metal-based photothermal catalyst with the nature of efficient light-to-thermal conversion and H2 dissociation, the wide-bandgap semiconductor needs to be modified to possess wide-wavelength-range absorption and the active surface. It remains a challenge to achieve the two aims simultaneously via single material modulation approach. In this study, one strategy of carbon doping can empower In2O3 with two advantageous modifications. The carbon doping can reduce the formation energy of oxygen vacancy, which induces the generation of oxygen-vacancy-riched material. The introduction of oxygen defect levels and carbon doping levels in band gap of In2O3 significantly reduces this band gap, which endows it full-spectral and intensive solar light absorption. Therefore, the carbon doped In2O3 achieves effective light-to-thermal conversion and delivers a 123.6 mmol g− 1 h− 1 of CO generation rate with near-unity selectivity, as well as prominent stability in photothermocatalytic CO2 reduction.

In this work, we propose a novel deep learning framework, called Attention-Based 2-layer Bi-ConvLSTM (denoted as Att-2BiConvLSTM) model, to predict the number of illegal-parking events in urban spaces. We model the research as a "next frame" prediction problem, which aims to improve urban transportation conditions and enhance the security and right-of-way for pedestrians. Various features in the prediction model are considered: some of them (e.g., hourly weather, traffic volumes) are dynamic every hour, while others (e.g., road network, point-of-interests) are static. To boost the effectiveness of static features, we propose a dynamic training process to transform the static features into dynamics. After that, all features can vary with time so that they are capable of handling a real-time prediction scenario. Moreover, we propose an attention mechanism for enhancing our bi-directional ConvLSTM model. With experimental verifications, we find that our proposed Att-2BiConvLSTM model can outperform other state-of-art and baseline methods. Besides, our model is useful for combining all features to make an accurate prediction.

Two-dimensional (2D) Janus materials are a novel kind of 2D materials, which have potential applications in nanoelectronics, optoelectronics and spintronics. However, a 2D Janus material combined with intrinsic ferromagnetism, electric dipole moment, valley polarization and Dzyaloshinskii–Moriya interaction (DMI) remains rarely reported. Here, the electronic structure and magnetic properties of 2D intrinsic ferromagnetic Janus 2H-VSeX (X = S, Te) monolayers are investigated systematically using the density-functional theory. Janus 2H-VSeX (X = S, Te) monolayers are intrinsic ferromagnetic semiconductors with in-plane magnetic anisotropy (IMA). The valley splitting of Janus 2H-VSeX (X = S, Te) monolayers appears by considering the spin–orbit coupling (SOC) effect and out of plane magnetization. Additionally, spontaneous vertical electric dipole moment and a large DMI are also found in Janus 2H-VSeX (X = S, Te) monolayers due to the broken inversion symmetry. Moreover, the valley splitting and DMI can be significantly increased by applying in-plane biaxial strain. These results provide an interesting 2D intrinsic ferromagnetic Janus material, which has potential applications in spintronic and valleytronic devices.

Software model checking is the technique that can automatically verify whether a given software meets the correctness properties. Although a large number of software model checkers have been developed, it remains a difficult task for practitioners to select a suitable checker for the software at hand, given the fact that the underlying techniques and performance trade-offs of these tools are hard to accurately characterize. In this paper, we study the algorithm selection problem for software model checking and apply the neural network techniques, in particular the neural networks with random weights (NNRW) to solve this problem. We also carry out a thorough experimental evaluation of the performance of three typical NNRW (i.e. RVFL, ELM, SCN) based on publicly available dataset. Our results demonstrate strong viability and usefulness of NNRW for this problem. To the best of our knowledge, this is the first work that applies NNRW techniques to algorithm selection for software model checking.

Activating transcription factor 2(ATF2), a transcription factor belonging to the AP-1 family, plays an important role in inflammation. However, its biological functions and underlying molecular mechanisms in rheumatoid arthritis (RA) remain unclear. Western blot and immunohistochemistry were used to identify the expression of ATF2 and Sprouty2(SPRY2) in RA synovial tissues. SW982 cells were stimulated by TNF-α to establish an in vitro RA fibroblast-like synoviocyte (RA-FLS) model. Transwell and monolayer wound-healing were used to detect cell migration and invasion. RNA interference (si-ATF2) and adenovirus vector (Ad-SPRY2) methods were employed to manipulate ATF2 or SPRY2 expression in SW982 cells. The protein expression and phosphorylation levels in SW982 cells were evaluated by western blot. ATF2 expression and phosphorylation were upregulated in the RA synovial tissues. In RA-FLS model, ATF2 expression and phosphorylation were increased in a time-dependent manner. ATF2 knockdown inhibited the migration and invasion of RA-FLS model, reducing the inflammatory factors, which was consistent with the influence on cell behaviors caused by SPRY2 overexpression. Moreover, SPRY2 overexpression inhibited the TNF-α-induced phosphorylation of ERK and ATF2 in SW982 cells. The high expression and phosphorylation of ATF2 promoted migration and invasion of RA-FLSs. SPRY2 might inhibited the inflammatory responses of RA-FLSs via suppressing ERK-ATF2 pathway.

The original version of this article contained mistakes, and the authors would like to correct them.

In a conventional stereo vision ranging system, two detectors placed horizontally calculate the distance between the target and the detector based on the triangular similarity principle, parallax and focal length. In general, the order of magnitude between the distance detected and the focal length varies widely. If there is an error in the parallax information, the calculated actual distance will be significantly different from the calculated distance. This paper proposes a new distance measuring principle that uses the vertical distance difference between two lenses placed along the shooting direction to eliminate errors caused by focal length. Finally, the distance between the target and the detector is calculated by the target image and the number of pixels occupied by the target.

Spark has extensively used in many applications of Tencent, due to its easy deployment, pipeline capability, and close integration with the Hadoop ecosystem. As the graph computing engine of Spark, GraphX is also widely deployed to process large-scale graph data in Tencent. However, when the size of the graph data is up to billion-scale, GraphX encounters serious performance degradation. Worse, Graphx cannot support the rising advancement of graph embedding (GE) and graph neural network (GNN) algorithms. To address these challenges, we develop a new graph processing system, called PSGraph, which uses Spark executor and PyTorch to perform calculation, and develops a distributed parameter server to store frequently accessed models. PSGraph can train extremely large-scale graph data in Tencent with the parameter server architecture, and enable the training of GE and GNN algorithms. Moreover, PSGraph still benefits from the advantages of Spark via staying inside the Spark ecosystem, and can directly replace GraphX without modification to the existing application framework. Our experiments show that PSGraph outperforms GraphX significantly.

Distributed machine learning (ML) has triggered tremendous research interest in recent years. Stochastic gradient descent (SGD) is one of the most popular algorithms for training ML models, and has been implemented in almost all distributed ML systems, such as Spark MLlib, Petuum, MXNet, and TensorFlow. However, current implementations often incur huge communication and memory overheads when it comes to large models. One important reason for this inefficiency is the row-oriented scheme (RowSGD) that existing systems use to partition the training data, which forces them to adopt a centralized model management strategy that leads to vast amount of data exchange over the network.We propose a novel, column-oriented scheme (ColumnSGD) that partitions training data by columns rather than by rows. As a result, ML model can be partitioned by columns as well, leading to a distributed configuration where individual data and model partitions can be collocated on the same machine. Following this locality property, we develop a simple yet powerful computation framework that significantly reduces communication overheads and memory footprints compared to RowSGD, for large-scale ML models such as generalized linear models (GLMs) and factorization machines (FMs). We implement ColumnSGD on top of Apache Spark, and study its performance both analytically and experimentally. Experimental results on both public and real-world datasets show that ColumnSGD is up to 930 x faster than MLlib, 63 x faster than Petuum, and 14 x faster than MXNet.

Holocene moisture evolution in the arid Central Asia region, dominated by the westerly circulation system, has been shown to be in drastic contrast with that in Asian monsoonal regions. Yet, water isotope records, including stalagmite oxygen isotopes and terrestrial long-chain n-alkane/acid hydrogen isotopes, show many common features in the two regions. Here we present several new isotopic records from the arid Central Asia region to examine the isotopic differences from various archives/media, together with existing water isotopic records from both regions. Isotopic records more reflecting terrestrial signal in arid regions appear to follow the pattern in monsoonal regions, while those likely affected by isotopic enrichment due to lake water evaporation display various patterns, and not necessarily resemble moisture changes inferred from the same lakes. It thus appears that the terrestrial water isotopes in both regions may record the isotopic signature in precipitation, but not necessarily linked to aridity changes. Meanwhile, those isotopic records affected by lake evaporation, after subtracting the original precipitation isotopic signal, show good correspondence to moisture changes.

The East Asian summer monsoon (EASM) system plays a crucial role in regulating climatic and hydrological variability in East Asia. However, it remains controversial how EASM altered the spatiotemporal patterns of hydrological variability of eastern China during the Little Ice Age (LIA). Here, we established a relatively high-resolution palaeohydrological record over the last 700 years in East China via analyzing n-alkanes from a peat core. Mid- and long-chain (C21–C25 and C27–C31) n-alkanes account for more than 95% of the total odd‑carbon-numbered n-alkanes, with short-chain n-alkanes (C17–C19) accounting for only 2.3% on average, suggesting that organic matter in the peat core originated primarily from vascular plants. Organic geochemical indices, including Paq (an indicator of contributions from aquatic plants), the average chain length (ACL) of long-chain n-alkanes, and C31/C27 and C23/C29 n-alkane ratios, suggest a transition from dry conditions during the middle LIA to wet conditions in the late LIA. Comparing Paq records from the study profile with previous records from the EASM region, we found that the northern and southern parts of eastern China showed contrasting changes in humidity during the LIA. Specifically, the southern part shifted from mid-LIA dry conditions to late-LIA wet conditions, whereas the northern part shifted from mid-LIA wet conditions to late-LIA dry conditions. This spatial heterogeneity in hydroclimate could be attributed to La Nina-like conditions and intense summer monsoons during the middle LIA and El Nino-like conditions and weak summer monsoons during the late LIA. Our study demonstrates the importance of understanding spatial heterogeneity of hydroclimatic changes to probe the mechanisms whereby EASM modulates regional climatic changes during the LIA.

The photoswitchable system with color and multi-state fluorescence changes is the key to the realization of advanced multi-level information storage, and its application potential has far exceeded the monochromatic fluorescent system. Herein, we introduced a novel multi-state photoswitchable fluorescent polymeric nanoparticles, which contains two energy-matched photoswitchable fluorescent molecules, i.e. diarylethene derivative (BH, donor) and spiropyran derivative (SPMA, acceptor), to construct a tunable fluorescence resonance energy transfer (FRET) system. By optimizing the feeding ratio of the two photochromic molecules to control the efficiency of energy transfer, the fluorescence of the nanoparticles can be reversibly switched among none, red, orange, and green by adjusting the wavelength of light stimuli. In addition, the nanoparticles display good water dispersibility, high brightness, fast responsiveness, excellent photoreversibility, and long-term fluorescence stability. We also verified that the system has great potential in optical data storage, multi-layer information encryption, and anti-counterfeiting.