Explore the vast range of titles available.

To address the challenges of label sparsity and feature incompleteness in structured data, a self-supervised representation learning method based on multi-view consistency constraints is proposed in this paper. Robust modeling of high-dimensional sparse tabular data is achieved through integration of a view-disentangled encoder, intra- and cross-view contrastive mechanisms, and a joint loss optimization module. The proposed method incorporates feature clustering-based view partitioning, multi-view consistency alignment, and masked reconstruction mechanisms, thereby enhancing the model’s representational capacity and generalization performance under weak supervision. Across multiple experiments conducted on four types of datasets, including user rating data, platform activity logs, and financial transactions, the proposed approach maintains superior performance even under extreme conditions of up to 40% feature missingness and only 10% label availability. The model achieves an accuracy of 0.87, F1-score of 0.83, and AUC of 0.90 while reducing the normalized mean squared error to 0.066. These results significantly outperform mainstream baseline models such as XGBoost, TabTransformer, and VIME, demonstrating the proposed method’s robustness and broad applicability across diverse real-world tasks. The findings suggest that the proposed method offers an efficient and reliable paradigm for modeling sparse structured data.

Binding of transcription factor (TF) proteins to regulatory DNA sites is key to accurate control of gene expression in response to environmental stimuli. Theoretical modeling of transcription regulation is often focused on a limited set of genes of interest, while binding of the TF to other genomic sites is seldom considered. The total number of TF binding sites (TFBSs) affects the availability of TF protein molecules and sequestration of a TF by TFBSs can promote bistability. For many signaling systems where a graded response is desirable for continuous control over the input range, biochemical parameters of the regulatory proteins need be tuned to avoid bistability. Here we analyze the mono-/bistable parameter range for positively autoregulated two-component systems (TCSs) in the presence of different numbers of competing TFBSs. TCS signaling, one of the major bacterial signaling strategies, couples signal perception with output responses via protein phosphorylation. For bistability, competition for TF proteins by TFBSs lowers the requirement for high fold change of the autoregulated transcription but demands high phosphorylation activities of TCS proteins. We show that bistability can be avoided with a low phosphorylation capacity of TCSs, a high TF affinity for the autoregulated promoter or a low fold change in signaling protein levels upon induction. These may represent general design rules for TCSs to ensure uniform graded responses. Examining the mono-/bistability parameter range allows qualitative prediction of steady-state responses, which are experimentally validated in the E . coli CusRS system.

An efficient and secure computation framework based on the sliding-window attention mechanism and sliding loss function was proposed to address challenges in temporal and spatial feature modeling for multimodal data processing. The framework aims to overcome the limitations of traditional methods in privacy protection, feature-capturing capabilities, and computational efficiency. The experimental results demonstrated that, in time-series data processing tasks, the proposed method achieved precision, recall, accuracy, and F1-score values of 0.95, 0.91, 0.93, and 0.93, respectively, significantly outperforming the federated learning, secure multi-party computation, homomorphic encryption, and TEE-based approaches. In spatial data processing tasks, these metrics reached 0.93, 0.90, 0.92, and 0.91, also surpassing all the comparative methods. Compared with the existing secure computation frameworks, the proposed approach substantially enhanced computational efficiency while minimizing accuracy loss, all while ensuring data privacy. These findings provide an efficient and reliable solution for privacy protection and data security in cloud computing environments. Furthermore, the research demonstrates significant theoretical value and practical potential in real-world scenarios such as financial forecasting and image analysis.

The uncertainty propagation of an orbit is crucial for ensuring the robustness of orbit design. In this paper, the lunar free return trajectory is selected as the research focus, and the uncertainty propagation under high-precision orbit prediction models is analyzed based on the self-developed Aerospace Tool Kit (ATK) software by the project team. Firstly, the design scheme for the nominal orbit of the lunar free return trajectory is provided. Secondly, a method for uncertainty propagation of the trajectory through Monte Carlo simulation is developed based on the secondary development of ATK. Finally, the uncertainty propagation of the lunar free return trajectory is achieved using ATK, and the results of the uncertainty propagation are preliminarily analyzed.

Li 7 La 3 Zr 2 O 12 gives rise to much attention in solid-state batteries due to the excellent ionic conductivity, stable contact with lithium metal, and wide electrochemical window. The Li 6.25 Al 0.25 La 3 Zr 2 O 12 electrolytes were prepared by using cold sintering and then annealed at different temperatures. The effects of liquid-phase medium and annealing temperatures on phase compositions, bulk densities, and ionic conductivities were investigated by XRD, SEM, and electrochemical impedance. With the increasing of annealing temperatures, the ionic conductivities of the samples increased firstly and then decreased as the binary Li 6.25 Al 0.25 La 3 Zr 2 O 12 phase and La 2 Zr 2 O 7 phase were formed. The ionic conductivity of 17.8 × 10 −3 mS/cm was obtained for the sample with LiOH solution additives, which were annealed at 1000 °C. It can be ascribed to the rearrangement of the grain particles with the help of LiOH liquid phase and the reduction in the free energy of particles. The cold sintering combined with the annealed process is an effective way to increase grain conductivity and promote the further study of solid-state electrolytes.

The present study aimed to develop and evaluate a nutritional and nursing risk assessment method for diabetic inpatients to improve healthcare and risk management. Diabetic inpatients diagnosed according to the World Health Organization guidelines, together with their nursing staff, were divided into two groups for nutritional and nursing risk assessment. Data from one group were used to establish the assessment method, and data from the other group were used to evaluate the reliability and effectiveness of the method. To establish the method, various risk variables in the nutritional and nursing processes were evaluated by logistic regression analysis; the score and probability of the risk variables were determined based on odds ratios. The overall nutritional and nursing risk for individual inpatients was then judged by the accumulated scores. The analysis showed that there were a number of risk factors, including age and body mass index. The risk was shown to increase with increasing score for the inpatients, and the χ2 test (P<0.01) was used to indicate a significant association. When the score was 50, the sensitivity and specificity of the method used to detect the nutritional and nursing risk were 88.3 and 66.5%, respectively, with predictive positive and negative rates of 12.83 and 98.53%, respectively. Therefore, the method is simple, cost-effective and fast; it can be used to screen a large number of patients by nursing staff and can also be used by patients themselves. Overall, the method is an effective and practicable nutritional and nursing risk assessment and educational tool.

Supported metal nanoparticles are of universal importance in many industrial catalytic processes. Unfortunately, deactivation of supported metal catalysts via thermally induced sintering is a major concern especially for high-temperature reactions. Here, we demonstrate that the particle distance as an inherent parameter plays a pivotal role in catalyst sintering. We employ carbon black supported platinum for the model study, in which the particle distance is well controlled by changing platinum loading and carbon black supports with varied surface areas. Accordingly, we quantify a critical particle distance of platinum nanoparticles on carbon supports, over which the sintering can be mitigated greatly up to 900 °C. Based on in-situ aberration-corrected high-angle annular dark-field scanning transmission electron and theoretical studies, we find that enlarging particle distance to over the critical distance suppress the particle coalescence, and the critical particle distance itself depends sensitively on the strength of metal-support interactions. Deactivation of supported metal catalysts via thermally induced sintering is a major concern in the catalysis community. Here, the authors demonstrate that enlarging particle distance to over the critical distance could suppress the particle coalescence greatly up to 900 °C.

The construction of functional three-dimensional covalent organic frameworks (3D COFs) for gas separation, specifically for the efficient removal of ethane (C 2 H 6 ) from ethylene (C 2 H 4 ), is significant but challenging due to their similar physicochemical properties. In this study, we demonstrate fine-tuning the pore environment of ultramicroporous 3D COFs to achieve efficient one-step C 2 H 4 purification. By choosing our previously reported 3D-TPB-COF-H as a reference material, we rationally design and synthesize an isostructural 3D COF (3D-TPP-COF) containing pyridine units. Impressively, compared with 3D-TPB-COF-H, 3D-TPP-COF exhibits both high C 2 H 6 adsorption capacity (110.4 cm 3 g −1 at 293 K and 1 bar) and good C 2 H 6 /C 2 H 4 selectivity (1.8), due to the formation of additional C-H···N interactions between pyridine groups and C 2 H 6 . To our knowledge, this performance surpasses all other reported COFs and is even comparable to some benchmark porous materials. In addition, dynamic breakthrough experiments reveal that 3D-TPP-COF can be used as a robust absorbent to produce high-purity C 2 H 4 directly from a C 2 H 6 /C 2 H 4 mixture. This study provides important guidance for the rational design of 3D COFs for efficient gas separation. The construction of three-dimensional covalent organic frameworks for gas separation is challenging due to the similar physicochemical properties of the gas mixture. Here, the authors report functional three-dimensional covalent organic frameworks by fine-tunning the pore environment with pyridine units to achieve effective separation of ethane from ethylene.

Background Although colonoscopy is the standard screening test for colorectal cancer (CRC), its use is limited by a poor compliance rate, the need for extensive bowel preparation, and the risk of complications. As an alternative, an FDA-approved stool-based DNA test, Cologuard, has demonstrated satisfactory detection performance for CRC, but its compliance rate remains suboptimal, primarily attributable to individuals’ reluctance to provide stool samples. Methods We developed a noninvasive blood-based CRC test, ColonSecure, based on cell-free DNA containing cancer-specific CpG island methylation patterns. We initially screened publicly available datasets for differentially methylated CpG sites in CRC with prediction potential. Subsequently, we performed two sequential bisulfite-free methylation sequencing on blood samples obtained from CRC patients and non-cancer controls. Through rigorous evaluation of each marker and machine learning-assisted feature selection, we identified 149 hypermethylated markers from over 193,000 CpG sites. These markers were then utilized to construct the ColonSecure model, enabling accurate CRC detection. Results We validated the efficacy of our cell-free DNA methylation-based blood test for CRC screening with 3493 high-risk individuals identified from 114,136 urban residents. The ColonSecure test identified 89 out of 103 CRC patients diagnosed by the follow-up colonoscopy, outperforming CEA, CRP, and CA19-9 (with a sensitivity of 86.4% compared to 45.6%, 39.8%, and 25.2% for CEA, CRP, and CA19-9 respectively; an AUROC of 0.956 compared to an AUROC of < 0.77 for other methods). Conclusion Our observations emphasize the potential of our multiple cfDNA methylation marker-based test for CRC screening in high-risk populations.

The large−scale integration of wind power and PV cells into electric grids alleviates the problem of an energy crisis. However, this is also responsible for technical and management problems in the power grid, such as power fluctuation, scheduling difficulties, and reliability reduction. The microgrid concept has been proposed to locally control and manage a cluster of local distributed energy resources (DERs) and loads. If the net load power can be accurately predicted, it is possible to schedule/optimize the operation of battery energy storage systems (BESSs) through economic dispatch to cover intermittent renewables. However, the load curve of the microgrid is highly affected by various external factors, resulting in large fluctuations, which makes the prediction problematic. This paper predicts the net electric load of the microgrid using a deep neural network to realize a reliable power supply as well as reduce the cost of power generation. Considering that the backpropagation (BP) neural network has a good approximation effect as well as a strong adaptation ability, the load prediction model of the BP deep neural network is established. However, there are some defects in the BP neural network, such as the prediction effect, which is not precise enough and easily falls into a locally optimal solution. Hence, a genetic algorithm (GA)−reinforced deep neural network is introduced. By optimizing the weight and threshold of the BP network, the deficiency of the BP neural network algorithm is improved so that the prediction effect is realized and optimized. The results reveal that the error reduction in the mean square error (MSE) of the GA–BP neural network prediction is 2.0221, which is significantly smaller than the 30.3493 of the BP neural network prediction. Additionally, the error reduction is 93.3%. The error reductions of the root mean square error (RMSE) and mean absolute error (MAE) are 74.18% and 51.2%, respectively.