Search

Synchronous federated learning (FL) over wireless networks often suffers from the straggler effect, when the time required for training local models and uploading trained parameters varies significantly across heterogeneous wireless devices. This disparity prolongs the duration needed for model aggregation at the data center and slows down the convergence of synchronous FL, posing a significant challenge for FL over wireless networks. In this paper, we propose a novel buffer-aided FL scheme to mitigate the straggler effect. A buffer with sufficiently large storage is deployed at each wireless device to temporarily store the collected training data and adaptively outputs it during local training, according to the computational capabilities and communication data rates of the wireless devices. Consequently, all local models can be synchronously aggregated at the data center to reduce the number of rounds required for model aggregation in FL. To ensure timely information updates, a staleness function is further introduced to characterize the freshness of the data used to train local models. Additionally, the entropic value-at-risk (EVaR) of the data queues is introduced to eliminate the impact of discarded data at the buffers and improve the accuracy of trained local models. We formulate a delay-aware online stochastic optimization problem to minimize the long-term average staleness of all wireless devices for buffer-aided FL. Our problem formulation simultaneously guarantees the stability of data queues at the wireless devices and reduces the risk of data loss. By employing the Lyapunov optimization technique, we transform the problem into instantaneous deterministic optimization subproblems and further solve each subproblem online via utilizing its hidden convexity. Simulation results demonstrate that the proposed buffer-aided synchronous FL scheme can effectively improve the convergence rate of FL and, at the same time, ensure timely synchronization of heterogeneous wireless devices.

The ocean is one of the essential fields of national defense in the future, and more and more attention is paid to the lightweight research of Marine equipment and materials. This study it is to develop a Machine learning (ML)-based prediction method to study the evolution of the mechanical properties of Al-Li alloys in the marine environment. We obtained the mechanical properties of Al-Li alloy samples under uniaxial tensile deformation at different exposure times through Marine exposure experiments. We obtained the strain evolution by digital image correlation (DIC). The strain field images are voxelized using 2D-Convolutional Neural Networks (CNN) autoencoders as input data for Long Short-Term Memory (LSTM) neural networks. Then, the output data of LSTM neural networks combined with corrosion features were input into the Back Propagation (BP) neural network to predict the mechanical properties of Al-Li alloys. The main conclusions are as follows: 1. The variation law of mechanical properties of 2297-T8 in the Marine atmosphere is revealed. With the increase in outdoor exposure test time, the tensile elastic model of 2297-T8 changes slowly, within 10%, and the tensile yield stress changes significantly, with a maximum attenuation of 23.6%. 2. The prediction model can predict the strain evolution and mechanical response simultaneously with an error of less than 5%. 3. This study shows that a CNN/LSTM system based on machine learning can be built to capture the corrosion characteristics of Marine exposure experiments. The results show that the relationship between corrosion characteristics and mechanical response can be predicted without considering the microstructure evolution of metal materials.

User preferences are important for electric vehicle charging station (EVCS) recommendations, but they have not been deeply analyzed. Therefore, in this study, user charging preferences are identified and applied to EVCS recommendations using a hybrid model that integrates LightGBM and singular value decomposition (SVD). In the model, LightGBM is used to predict user ratings according to users’ comments regarding charging orders, and the feature importance reported by each user is output. Then, a co-occurrence matrix between users and charging stations (EVCSs) is constructed and decomposed using SVD. Based on the decomposed results, the final evaluated scores of each user for EVCSs can be calculated. Upon ranking the EVCSs according to the scores, the EVCS recommendation results are obtained, taking into account the users’ charging preferences. The sample data consist of 28,306 orders from 508 users at 241 charging stations in Linyi, Shandong, China. The experimental results show that the proposed hybrid model outperforms the benchmark models in terms of precision, recall, and F1 score, and its F1 score can be increased by 96% compared with that of the traditional item-based collaborative filtering method with charging counts for EVCS recommendations.

Bone defects caused by diseases or surgery are a common clinical problem. Researchers are devoted to finding biological mechanisms that accelerate bone defect repair, which is a complex and continuous process controlled by many factors. As members of transcriptional costimulatory molecules, Yes-associated protein (YAP) and transcriptional co-activator with PDZ-binding motif (TAZ) play an important regulatory role in osteogenesis, and they affect cell function by regulating the expression of osteogenic genes in osteogenesis-related cells. Macrophages are an important group of cells whose function is regulated by YAP/TAZ. Currently, the relationship between YAP/TAZ and macrophage polarization has attracted increasing attention. In bone tissue, YAP/TAZ can realize diverse osteogenic regulation by mediating macrophage polarization. Macrophages polarize into M1 and M2 phenotypes under different stimuli. M1 macrophages dominate the inflammatory response by releasing a number of inflammatory mediators in the early phase of bone defect repair, while massive aggregation of M2 macrophages is beneficial for inflammation resolution and tissue repair, as they secrete many anti-inflammatory and osteogenesis-related cytokines. The mechanism of YAP/TAZ-mediated macrophage polarization during osteogenesis warrants further study and it is likely to be a promising strategy for bone defect repair. In this article, we review the effect of Hippo-YAP/TAZ signaling and macrophage polarization on bone defect repair, and highlight the regulation of macrophage polarization by YAP/TAZ.

Osteoporosis （OP） is a systemic bone disease and a common orthopedic disorder in elderly patients. High morbidity， disability and mortality rates and medical cost of OP bring huge burden to the patients and their family. Serine/threonine protein kinase（AMPK）/ mammalian target of rapamycin （mTOR） signaling pathway plays a significant role in cell metabolism and apoptosis. Traditional Chinese medicine monomer or compound can prevent and treat OP by regulating the expression levels of related factors through the AMPK/mTOR signaling pathway， whereas the specific mechanism remains elusive. In this article， the effect of AMPK/mTOR signaling pathway upon osteoblasts and osteoclasts， and the regulatory role of traditional Chinese medicine were reviewed， aiming to provide novel perspectives and ideas for prevention and treatment of OP.

Abstract Background Venous thromboembolism (VTE) is a common postoperative complication in patients undergoing surgery for gastric cancer (GC). Although VTE incidence may vary among cancers, guidelines rarely stratify preventive methods for postoperative VTE by cancer type. The risk of VTE in patients undergoing surgery for GC remains unclear. Methods A systematic review and meta-analysis was undertaken to determine the risk of VTE after GC surgery and discuss the clinical value of pharmacological thromboprophylaxis in these cases. Medline, Embase, Web of Science, and Cochrane Library databases were searched for articles published from their inception to September 2022. Results Overall, 13 studies (111,936 patients) were included. The overall 1-month incidence of VTE, deep vein thrombosis (DVT), and pulmonary embolism (PE) after GC surgery was 1.8% (95% CI, 0.8–3.1%; I²=98.5%), 1.2% (95% CI, 0.5–2.1%; I²=96.1%), and 0.4% (95% CI, 0.1–1.1%; I²=96.3%), respectively. The prevalence of postoperative VTE was comparable between Asian and Western populations (1.8% vs. 1.8%; P > 0.05). Compared with mechanical prophylaxis alone, mechanical plus pharmacological prophylaxis was associated with a significantly lower 1-month rate of postoperative VTE and DVT (0.6% vs. 2.9% and 0.6% vs. 2.8%, respectively; all P 0.05). The 1-month postoperative incidence of VTE was not significantly different between laparoscopic and open surgery (1.8% vs. 4.3%, P > 0.05). Conclusion Patients undergoing GC surgery do not have a high risk of VTE. The incidence of VTE after GC surgery is not significantly different between Eastern and Western patients. Mechanical plus pharmacological prophylaxis is more effective than mechanical prophylaxis alone in postoperative VTE prevention. The VTE risk is comparable between open and laparoscopic surgery for GC.

Complex systems (CSs) have garnered increasing attention in science education due to their prevalence in the natural world and their importance in addressing pressing issues such as climate change, pandemics, and biodiversity loss. However, the instruments for assessing one’s CS knowledge are limited, and the knowledge of CSs among in-service teachers remains underreported. Guided by the complexity framework, we launched a study to develop and validate a survey instrument for measuring the knowledge of CSs among high school teachers and undergraduate students, as well as delineating the contours of their knowledge. In this article, we present the development and validation of a Complex Systems Knowledge Survey (CSKS), and we use the survey to compare the CS knowledge among 252 high school teachers and 418 undergraduate students in the United States. Our key findings include that (1) both high school teachers and undergraduates exhibit relatively low knowledge of decentralization and stochasticity, (2) undergraduates, especially those in non-STEM majors, demonstrate moderate to low knowledge of emergence, (3) few teachers and undergraduates differentiate between complicated and complex systems, and (4) teachers and undergraduates recognize CS examples across natural and social systems. The implications of our findings are discussed.