Search

The steel–aluminum hybrid body closure panels can achieve a more balanced and lightweight performance. However, the differences in the physical properties of metal sheets and the complex changes in the properties of the adhesive material result in cumulative deviations in the composite-forming process. This paper proposes a deformation pre-compensation modeling method for the autobody closure panels hemming system oriented towards the process chain, in response to the problem that single-process optimization cannot obtain global optimal results. Taking the car door scaled model as an example, based on surface reconstruction and node compensation, the curing deformation amount is fed back in advance to the gluing and hemming processes. The deformation deviation is corrected through geometric parameter pre-compensation to achieve overall process shape control and optimization. Research shows that this method can significantly reduce the surface differences and gaps of hemming structures with dissimilar materials, and a single iteration can reduce the assembly surface difference by more than 90%. This provides a reference for improving the manufacturing quality of steel–aluminum hybrid body closure panels. [ABSTRACT FROM AUTHOR]

Rapid urbanization has altered the natural surface properties and spatial patterns, increasing the risk of urban waterlogging. Assessing the probability of urban waterlogging risk is crucial for preventing and mitigating the environmental risks associated with urban waterlogging. This study aims to evaluate the impact of different urban spatial morphologies on the probability of urban waterlogging risk. The proposed assessment framework was demonstrated in Guangzhou, a high-density city in China. Firstly, a spatial weight naive Bayes model was employed to map the probability of waterlogging risk in Guangzhou. Secondly, the World Urban Database and Access Portal Tools (WUDAPT)-based method was used to create a local climate zone (LCZ) map of Guangzhou. Then, the range of waterlogging risk and the proportion of risk levels were analyzed across different LCZs. Finally, the Theil index was used to measure the disparity in waterlogging risk exposure among urban residents. The results indicate that 16.29% of the area in Guangzhou is at risk of waterlogging. Specifically, 13.06% of the area in LCZ 2 is classified as high risk, followed by LCZ 1, LCZ 8, and LCZ 10, with area proportions of 11.42%, 8.37%, and 6.26%, respectively. Liwan District has the highest flood exposure level at 0.975, followed by Haizhu, Yuexiu, and Baiyun. The overall disparity in waterlogging exposure in Guangzhou is 0.30, with the difference between administrative districts (0.13) being smaller than the difference within the administrative districts (0.17). These findings provide valuable insights for future flood risk mitigation and help in adopting effective risk reduction strategies at urban planning level. [ABSTRACT FROM AUTHOR]

Counting the number of trees and obtaining information on tree crowns have always played important roles in the efficient and high-precision monitoring of forest resources. However, determining how to obtain the above information at a low cost and with high accuracy has always been a topic of great concern. Using deep learning methods to segment individual tree crowns in mixed broadleaf forests is a cost-effective approach to forest resource assessment. Existing crown segmentation algorithms primarily focus on discrete trees, with limited research on mixed broadleaf forests. The lack of datasets has resulted in poor segmentation performance, and occlusions in broadleaf forest images hinder accurate segmentation. To address these challenges, this study proposes a supervised segmentation method, SegcaNet, which can efficiently extract tree crowns from UAV images under natural light conditions. A dataset for dense mixed broadleaf forest crown segmentation is produced, containing 18,000 single-tree crown images and 1200 mixed broadleaf forest images. SegcaNet achieves superior segmentation results by incorporating a convolutional attention mechanism and a memory module. The experimental results indicate that SegcaNet's m I o U values surpass those of traditional algorithms. Compared with FCN, Deeplabv3, and MemoryNetV2, SegcaNet's m I o U is increased by 4.8%, 4.33%, and 2.13%, respectively. Additionally, it reduces instances of incorrect segmentation and over-segmentation. [ABSTRACT FROM AUTHOR]

Polyhalogenated carbazoles (PHCZs) are a class of nitrogen-containing heterocyclic compounds that are widely distributed throughout the marine environment and sediment. These compounds share structural and toxicity similarities with dioxins. However, our understanding of the toxicological effects of PHCZs on marine organisms and their underlying molecular mechanisms remains limited. In this study, we employed the marine model organism Mugilogobius chulae as the experimental subject and selected 2,7-dibromocarbazole (2,7-DBCZ), a compound known for its high toxicity and detection frequency, to conduct both an acute toxicity test and transcriptome analysis on M. chulae embryos. Our findings revealed that the 96 h median lethal concentration (LC50) of 2,7-DBCZ for M. chulae embryos was 174 μg/L, with a median effective concentration (EC50) resulting in pericardial edema deformity of 88.82 μg/L. Transcriptome analysis revealed significant impacts on various systems in M. chulae embryos following exposure to 2,7-DBCZ, including the sensory, cardiovascular, immune, and endocrine systems. Furthermore, this compound perturbed signaling pathways such as phototransduction, protein folding and processing, amino acid metabolism, lipid transport, and exogenous compound metabolism. Notably, transcript abundance of the CYP1A gene associated with the activation of the AhR signaling pathway, similar to dioxin-like compounds, was 18.18 times higher than that in the control group. This observation suggests that M. chulae embryos mount a stress response when exposed to PHCZs. In summary, this study contributes to our understanding of the toxicological implications of PHCZ in marine fish and offers a theoretical foundation for risk assessment and regulatory frameworks for PHCZs in the marine environment. [ABSTRACT FROM AUTHOR]

Few-shot learning (FSL) is a challenging problem. Transfer learning methods offer a straightforward and effective solution to FSL by leveraging pre-trained models and generalizing them to new tasks. However, pre-trained models often lack the ability to highlight and emphasize salient features, a gap that attention mechanisms can fill. Unfortunately, existing attention mechanisms encounter issues such as high complexity and incomplete attention information. To address these issues, we propose a dimensionally enhanced attention (DEA) module for FSL. This DEA module introduces minimal additional computational overhead while fully attending to both channel and spatial information. Specifically, the feature map is first decomposed into 1D tensors of varying dimensions using strip pooling. Next, a multi-dimensional collaborative learning strategy is introduced, enabling cross-dimensional information interactions through 1D convolutions with adaptive kernel sizes. Finally, the feature representation is enhanced by calculating attention weights for each dimension using a sigmoid function and weighting the original input accordingly. This approach ensures comprehensive attention to different dimensions of information, effectively characterizing data in various directions. Additionally, we have found that knowledge distillation significantly improves FSL performance. To this end, we implement a logit standardization self-distillation method tailored for FSL. This method addresses the issue of exact logit matching, which arises from the shared temperature in the self-distillation process, by employing logit standardization. We present experimental results on several benchmark datasets where the proposed method yields significant performance improvements. [ABSTRACT FROM AUTHOR]

Masson pine (Pinus massoniana Lamb.) is a tree species that is widely distributed throughout southern China and holds significant economic and ecological value. The main objective of our study was to assess the effects of thinning on aboveground biomass increments and tree diversity in both the overstory and understory. Additionally, the underlying factors and mechanisms responsible for driving changes in biomass increment were analyzed. Four different thinning treatments (control, light thinning, moderate thinning, and heavy thinning) were implemented in 214 plots (~1800 tree ha−1) in three Masson pine forests in Hunan Province, China. A robustly designed experiment was used with over six years of repeated measurements. The differences in biomass increment and tree diversity among the different treatments were compared using repeated measures ANOVAs. The Mantel test was used to determine environmental metrics correlated with biomass increments across tree strata. Structural equation modeling was utilized to explore the multivariate relationships among site environment, tree diversity, and post-treatment biomass increment. The results indicated that thinning overall increased biomass increment, the Shannon index, and the Gini index, while decreasing the Dominance index over time. Moderate thinning (25%–35% of trees removed) was found to promote overstory biomass increment to 9.72 Mg·ha−1·a−1 and understory biomass increment to 1.43 Mg·ha−1·a−1 six years post-thinning, which is significantly higher than that of other treatments. Environmental metrics such as light intensity, soil organic matter, and other soil physiochemical properties were positively correlated with biomass increments, and their effects on the overstory and understory differed. Structural equation modeling revealed that thinning treatments, environmental metrics, tree diversity, and their interactions could be the main drivers for biomass increments across tree strata. Specifically, thinning treatments, light intensity, and tree size diversity (Gini index) had significant effects on overstory biomass increment, while understory species richness (Shannon index) and soil organic matter affected understory biomass increment. In conclusion, moderate thinning is an effective silvicultural treatment for stimulating biomass increments of both the overstory and understory in Masson pine forests in southern China if a middle period (e.g., six years) is considered. Some factors, such as species richness, tree size diversity, and environmental metrics (e.g., light and soil), are suggested for consideration to improve the efficiency of thinning. [ABSTRACT FROM AUTHOR]

Urbanization and climate change have led to rising urban temperatures, increasing heat-related health risks. Assessing urban heat risk is crucial for understanding and mitigating these risks. Many studies often overlook the impact of block types on heat risk, which limits the development of mitigation strategies during urban planning. This study aims to investigate the influence of various spatial factors on the heat risk at the block scale. Firstly, a GIS approach was used to generate a Local Climate Zones (LCZ) map, which represents different block types. Secondly, a heat risk assessment model was developed using hazard, exposure, and vulnerability indicators. Thirdly, the risk model was demonstrated in Guangzhou, a high-density city in China, to investigate the distribution of heat risk among different block types. An XGBoost model was used to analyze the impact of various urban spatial factors on heat risk. Results revealed significant variations in heat risk susceptibility among different block types. Specifically, 33.9% of LCZ 1–4 areas were classified as being at a high-risk level, while only 23.8% of LCZ 6–9 areas fell into this level. In addition, the pervious surface fraction (PSF) had the strongest influence on heat risk level, followed by the height of roughness elements (HRE), building surface fraction (BSF), and sky view factor (SVF). SVF and PSF had a negative impact on heat risk, while HRE and BSF had a positive effect. The heat risk assessment model provides valuable insights into the spatial characteristics of heat risk influenced by different urban morphologies. This study will assist in formulating reasonable risk mitigation measures at the planning level in the future. [ABSTRACT FROM AUTHOR]

In this study, a multi-stage planning framework was constructed by using SWMM simulation modeling and NSGA-II and applied to optimize the layout of integrated grey–green infrastructure (IGGI) under land use change and climate change scenarios. The land use change scenarios were determined based on the master plan of the study area, with imperviousness of 50.7% and 62.0% for stage 1 and stage 2, respectively. Rainfall trends for stage 1 and stage 2 were determined using Earth-E3 from the CMIP6 model. The rainfall in stage 2 increased by 14.9% from stage 1. Based on these two change scenarios, the spatial configuration of IGGI layouts with different degrees of centralization of the layout (DCL) under the two phases was optimized, with the lowest life cycle cost (LCC) as the optimization objective. The results showed that the layout with DCL = 0 had better performance in terms of LCC. The LCC of the layout with DCL = 0 was only 66.9% of that of the layout with DCL = 90.9%. In terms of Tech-R, stage 2 had better performance than stage 1. Furthermore, the average technological resilience (Tech-R) index of stage 2 was 0.8–3.4% higher than that of stage 1. Based on the LCC and Tech-R indices of all of the layouts, TOPSIS was used to compare the performance of the layouts under the two stages, and it was determined that the layout with DCL = 0 had the best economic and performance benefits. The results of this study will be useful in exploring the spatial configuration of urban drainage systems under land use change and climate change for sustainable stormwater management. [ABSTRACT FROM AUTHOR]

Electric power operation, as one of the key fields in the world, faces particularly prominent safety issues. Ensuring the safety of operators has become the most fundamental requirement in power operation. However, there are some safety hazards in power construction. These hazards are mainly due to weak safety awareness among staff and the failure to standardize the wearing of safety helmets. In order to effectively address this situation, technical means such as video surveillance technology and computer vision technology can be utilized to monitor whether staff are wearing helmets and provide timely feedback. Such measures will greatly enhance the safety level of power operation. This paper proposes an improved lightweight helmet detection algorithm named YOLO-M3C. The algorithm first replaces the YOLOv5s backbone network with MobileNetV3, successfully reducing the model size from 13.7 MB to 10.2 MB, thereby increasing the model's detection speed from 42.0 frames per second to 55.6 frames per second. Then, the CA attention mechanism is introduced into the backbone network to enhance the feature extraction capability of the model. Finally, in order to further improve the detection recall rate and accuracy of the model, a knowledge distillation of the model was carried out. The experimental results show that, compared with the original YOLOv5s algorithm, the average accuracy of the improved YOLO-M3C algorithm is improved by 0.123, and the recall rate is the same. These results verify that the algorithm YOLO-M3C has excellent performance in target detection and recognition, which can improve accuracy and confidence, while reducing false detection and missing detection, and effectively meet the needs of helmet-wearing detection. [ABSTRACT FROM AUTHOR]

Skyrise greenery, including green roofs and vertical gardens, has emerged as an indispensable tool for sustainable urban planning with multiple ecological and economic benefits. A bibliometric analysis was used to provide a systematic review of the functions associated with skyrise greenery in urban landscapes. Key research tools, including the "Bibliometrix" R package and "CiteSpace" 6.2 R4, highlight the depth and breadth of the literature covering skyrise greenery. In 2000–2022, a total of 1474 original journal articles were retrieved. Over this period, there was an exponential increase in the number of publications, reflecting both enhanced knowledge and increasing concerns regarding climate change, the urban heat island, and urbanization. Of the total, ~58% of the articles originated from China, followed by the USA, Italy, Australia, and Canada. The research themes, such as urban heat islands, carbon sequestration, hydrology, and air quality, have been identified as the frontier in this fields. Furthermore, researchers from developed countries contributed the most publications to this domain, while developing countries, such as China, play an increasing role in the design and performance evaluation of vertical greenery. Key benefits identified in vertical green systems (e.g., green roofs and walls) include thermal regulation, sustainable water management, air-quality improvement, noise reduction, and biodiversity enhancement. In addition, several potential future research prospectives are highlighted. This review provides a comprehensive insight into exploring the pivotal role of skyrise greenery in shaping sustainable, resilient urban futures, coupled with sustainable urban planning. [ABSTRACT FROM AUTHOR]

Computer-aided diagnosis systems play a crucial role in the diagnosis and early detection of breast cancer. However, most current methods focus primarily on the dual-view analysis of a single breast, thereby neglecting the potentially valuable information between bilateral mammograms. In this paper, we propose a Four-View Correlation and Contrastive Joint Learning Network (FV-Net) for the classification of bilateral mammogram images. Specifically, FV-Net focuses on extracting and matching features across the four views of bilateral mammograms while maximizing both their similarities and dissimilarities. Through the Cross-Mammogram Dual-Pathway Attention Module, feature matching between bilateral mammogram views is achieved, capturing the consistency and complementary features across mammograms and effectively reducing feature misalignment. In the reconstituted feature maps derived from bilateral mammograms, the Bilateral-Mammogram Contrastive Joint Learning module performs associative contrastive learning on positive and negative sample pairs within each local region. This aims to maximize the correlation between similar local features and enhance the differentiation between dissimilar features across the bilateral mammogram representations. Our experimental results on a test set comprising 20% of the combined Mini-DDSM and Vindr-mamo datasets, as well as on the INbreast dataset, show that our model exhibits superior performance in breast cancer classification compared to competing methods. [ABSTRACT FROM AUTHOR]

The dimensional accuracy and microstructure affect the service performance of parts fabricated by wire arc additive manufacturing (WAAM). Regulating the geometry and microstructure of such parts presents a challenge. The coupling method of an artificial neural network and finite element (FE) is proposed in this research for this purpose. Back-propagating neural networks (BPNN) based on optimization algorithms were established to predict the bead width (BW) and height (BH) of the deposited layers. Then, the bead geometry was modeled based on the predicted dimension, and 3D FE heat transfer simulation was performed to investigate the evolution of temperature and microstructure. The results showed that the errors in BW and BH were less than 6%, and the beetle antenna search BPNN model had the highest prediction accuracy compared to the other models. The simulated melt pool error was less than 5% with the experimental results. The decrease in the ratio of the temperature gradient and solidification rate induced the transition of solidified grains from cellular crystals to columnar dendrites and then to equiaxed dendrites. Accelerating the cooling rate increased the primary dendrite arm spacing and δ-ferrite content. These results indicate that the coupling model provides a pathway for regulating the dimensions and microstructures of manufactured parts. [ABSTRACT FROM AUTHOR]

Nature-based solutions (NbSs) are considered to form an innovative stormwater management approach that has living resolutions grounded in natural processes and structures. NbSs offer many other environmental benefits over traditional grey infrastructure, including reduced air pollution and climate change mitigation. This review predominantly centers on the hydrological aspect of NbSs and furnishes a condensed summary of the collective understanding about NbSs as an alternatives for stormwater management. In this study, which employed the CIMO (Context, Intervention, Mechanism, Outcome) framework, a corpus of 187 NbS-related publications (2000–2023) extracted from the Web of Science database were used, and we expounded upon the origins, objectives, and significance of NbSs in urban runoff and climate change, and the operational mechanisms of NbSs (including green roofs, permeable pavements, bioretention systems, and constructed wetlands), which are widely used in urban stormwater management, were also discussed. Additionally, the efficacy of NbSs in improving stormwater quality and quantity is discussed in depth in this study. In particular, the critical role of NbSs in reducing nutrients such as TSS, TN, TP, and COD and heavy metal pollutants such as Fe, Cu, Pb, and Zn is emphasized. Finally, the main barriers encountered in the promotion and application of NbSs in different countries and regions, including financial, technological and physical, regulatory, and public awareness, are listed, and future directions for improving and strategizing NbS implementation are proposed. This review gathered knowledge from diverse sources to provide an overview of NbSs, enhancing the comprehension of their mechanisms and applications. It underscores specific areas requiring future research attention. [ABSTRACT FROM AUTHOR]

The notable absence of juvenile Pseudolarix amabilis trees in forest understories suggests their vulnerability to ecological niche competition, leading to limited survival prospects. This study examines the key factors limiting the growth of P. amabilis seedlings by investigating the effects of five ecological factors: light intensity, rainfall, groundwater level, soil type, and type of fertilization, on the growth of one-year-old P. amabilis seedlings. Our results demonstrate that increasing the light intensity promotes plant growth by augmenting the leaf count, leaf biomass, plant height, stem biomass, root biomass, and total biomass. Further analysis reveals that increased light intensity influences biomass allocation, reducing the specific leaf area and leaf–stem biomass ratio, and favoring root and stem growth over leaf investment. Rainfall, groundwater level, fertilization type, and rhizosphere soil type primarily influence root growth by impacting the soil's physicochemical properties. Specifically, rising groundwater levels lower the soil temperature and increase the soil moisture, total potassium content, and soil pH, leading to reductions in root biomass, plant height, net height increment, leaf number, and total biomass. When groundwater levels reach 21 cm and 28 cm, submerging the surface soil layer, root biomass decreases by 1.6 g/plant (−51.6%) and 2.3 g/plant (−74.2%), respectively. Further analysis reveals a gradual decrease in the root–shoot ratio above the 14 cm groundwater level, while the specific leaf area and leaf–stem biomass ratio remains unaffected, indicating stronger belowground root stress compared to aboveground stem and leaf components. The results highlight light intensity as the key ecological factor determining the growth of P. amabilis seedlings. These findings underscore the importance of considering light intensity in the management of natural stands, the cultivation of artificial forests, and the nursery cultivation of endangered P. amabilis. [ABSTRACT FROM AUTHOR]

Extensive research has shown that nanolayered structures are capable of suppressing the shear banding in metallic glass in nanoindentation experiments. However, the specific mode and mechanism of the shear banding underneath the indenter remains unknown. Also, the quantification of shear banding-induced strain localization is still a challenge. Herein, the size-dependent shear banding behavior of a CuTiZrNb high-entropy alloy-based nanolayered glass with individual layer thicknesses (h) ranging from 5 to 80 nm was systematically investigated by nanoindentation tests. It was found that the hardness of the designed structure was almost size-independent. Yet, a clear transition in the deformation modes from the cutting-like shear bands to the kinking-like ones was discovered as h decreased to 10 nm. Moreover, multiple secondary shear bands also appeared, in addition to the primary ones, in the sample with h = 10 nm. The transition leads to an obvious strain delocalization, as clearly illustrated by the proposed theoretical model, which is based on the assumption of a pure shear stress state to quantify the shear banding-induced strain localization. The strain delocalization results from the higher density of amorphous/amorphous interfaces that exhibit the change in morphology with a refined layered glass structure. [ABSTRACT FROM AUTHOR]

Concurrent meteorological extremes (CMEs) represent a class of pernicious climatic events characterized by the coexistence of two extreme weather phenomena. Specifically, the juxtaposition of Urban Extreme Rainfall (UER) and Urban Extreme Heat (UEH) can precipitate disproportionately deleterious impacts on both ecological systems and human well-being. In this investigation, we embarked on a meticulous risk appraisal of CMEs within China's Greater Bay Area (GBA), harnessing the predictive capabilities of three shared socioeconomic pathways (SSPs) namely, SSP1-2.6, SSP3-7.0, and SSP5-8.5, in conjunction with the EC-Earth3-Veg-LR model from the CMIP6 suite. The findings evidence a pronounced augmentation in CME occurrences, most notably under the SSP1-2.6 trajectory. Intriguingly, the SSP5-8.5 pathway, typified by elevated levels of greenhouse gas effluents, prognosticated the most intense CMEs, albeit with a temperate surge upon occurrence. Additionally, an ascendant trend in the ratio of CMEs to the aggregate of UER and UEH portends an escalating susceptibility to these combined events in ensuing decades. A sensitivity analysis accentuated the pivotal interplay between UER and UEH as a catalyst for the proliferation of CMEs, modulated by alterations in their respective marginal distributions. Such revelations accentuate the imperative of assimilating intricate interdependencies among climatic anomalies into evaluative paradigms for devising efficacious climate change countermeasures. The risk assessment paradigm proffered herein furnishes a formidable instrument for gauging the calamitous potential of CMEs in a dynamically shifting climate, thereby refining the precision of prospective risk estimations. [ABSTRACT FROM AUTHOR]

Simple Summary: The effect of macrophages on tumor cells in the tumor microenvironment of hepatocellular carcinoma (HCC) has attracted more attention. In this study, expression levels of MMP9 and MICA were significantly elevated in macrophages and tumor cells, respectively. The IFN-γ pathway was activated in MICA+ tumor cells and MMP9+ macrophages. The interaction between MICA+ tumor cells and MMP9+ macrophages was mediated through the PROS1-AXL pathway. MMP9 was mainly expressed in M2-like macrophages. IRF1 induced the upregulation of PROS1 and MICA in HCC cells. MICA+ tumor cells stimulated the secretion of MMP9 from macrophages through the PROS1-AXL axis, thereby facilitating the proteolytic shedding of MICA into soluble MICA. This study provided valuable insights and supported the therapeutic application of AXL inhibition in HCC. Background: tumor-associated macrophages (TAMs) constitute a significant proportion of non-cancerous cells within the intricate tumor microenvironment (TME) of hepatocellular carcinoma (HCC). Understanding the communication between macrophages and tumor cells, as well as investigating potential signaling pathways, holds promise for enhancing therapeutic responses in HCC. Methods: single-cell RNA-sequencing data and bulk RNA-sequencing data were derived from open source databases Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA). Through this analysis, we elucidated the interactions between MICA+ tumor cells and MMP9+ macrophages, primarily mediated via the PROS1-AXL axis in advanced HCC. Subsequently, we employed a range of experimental techniques including lentivirus infection, recombinant protein stimulation, and AXL inhibition experiments to validate these interactions and unravel the underlying mechanisms. Results: we presented a single-cell atlas of advanced HCC, highlighting the expression patterns of MICA and MMP9 in tumor cells and macrophages, respectively. Activation of the interferon gamma (IFN-γ) signaling pathway was observed in MICA+ tumor cells and MMP9+ macrophages. We identified the existence of an interaction between MICA+ tumor cells and MMP9+ macrophages mediated via the PROS1-AXL axis. Additionally, we found MMP9+ macrophages had a positive correlation with M2-like macrophages. Subsequently, experiments validated that DNA damage not only induced MICA expression in tumor cells via IRF1, but also upregulated PROS1 levels in HCC cells, stimulating macrophages to secrete MMP9. Consequently, MMP9 led to the proteolysis of MICA. Conclusion: MICA+ HCC cells secreted PROS1, which upregulated MMP9 expression in macrophages through AXL receptors. The increased MMP9 activity resulted in the proteolytic shedding of MICA, leading to the release of soluble MICA (sMICA) and the subsequent facilitation of tumor immune escape. [ABSTRACT FROM AUTHOR]

Postprandial Hyperglycemia (PPHG) persistently threatens patients' health. Therefore, accurate diabetes prediction is crucial for effective blood glucose management. Most current methods primarily focus on analyzing univariate blood glucose data using traditional neural networks, neglecting the importance of spatiotemporal modeling of multivariate data at the node and subgraph levels. This study aimed to evaluate the accuracy of using deep learning (DL) techniques to predict diabetes based on multivariable blood glucose data, aiming to improve resource allocation and decision-making in healthcare. We introduce a Nonlinear Aggregated Graph Neural Network (NLAGNN) that utilizes continuous multivariate historical blood glucose data from multiple patients to predict blood glucose levels over time, addressing the challenge of accurately extracting strong and weak correlation features. We preliminarily propose a Nonlinear Fourier Graph Neural Operator (NFGO) for nonlinear node representation, which effectively reduces meaningless noise. Additionally, a dynamic partitioning of graphs is introduced, which divides the a hypergraph into distinct subgraphs, enabling the further processing of strongly correlated features at the node and subgraph levels, ultimately obtaining the final prediction through layer aggregation. Extensive experiments on three datasets show that our proposed method achieves competitive results compared to existing advanced methods. [ABSTRACT FROM AUTHOR]

Computer-aided diagnosis systems play a crucial role in the diagnosis and early detection of breast cancer. However, most current methods focus primarily on the dual-view analysis of a single breast, thereby neglecting the potentially valuable information between bilateral mammograms. In this paper, we propose a Four-View Correlation and Contrastive Joint Learning Network (FV-Net) for the classification of bilateral mammogram images. Specifically, FV-Net focuses on extracting and matching features across the four views of bilateral mammograms while maximizing both their similarities and dissimilarities. Through the Cross-Mammogram Dual-Pathway Attention Module, feature matching between bilateral mammogram views is achieved, capturing the consistency and complementary features across mammograms and effectively reducing feature misalignment. In the reconstituted feature maps derived from bilateral mammograms, the Bilateral-Mammogram Contrastive Joint Learning module performs associative contrastive learning on positive and negative sample pairs within each local region. This aims to maximize the correlation between similar local features and enhance the differentiation between dissimilar features across the bilateral mammogram representations. Our experimental results on a test set comprising 20% of the combined Mini-DDSM and Vindr-mamo datasets, as well as on the INbreast dataset, show that our model exhibits superior performance in breast cancer classification compared to competing methods.