新疆财经大学

Attachment style, as a relatively stable personality trait, plays a significant role in explaining employees’ psychological and behavioral responses in the workplace. Drawing on conservation of resources theory, this study investigates the mechanisms through which employee attachment styles influence emotional exhaustion, including a potential mediator and the boundary conditions moderating these effects. Quantitative data were collected from 338 employees using a two-wave survey of employees across various industries in China. SPSS and AMOS were used for data analysis and hypothesis testing. The analysis reveals that both attachment avoidance and attachment anxiety significantly exacerbate emotional exhaustion through the mediation of leader–member exchange. Furthermore, a supportive organizational climate mitigates the impact of leader–member exchange on emotional exhaustion and weakens the indirect effect of attachment style on emotional exhaustion through leader–member exchange. These findings provide nuanced insights into the interplay among attachment styles, workplace relationships, and employee well-being in the specific organizational environment of China. The study also offers actionable recommendations for organizations to promote employees’ psychological health and work quality, including fostering supportive organizational climates and enhancing leader–member exchange.

Studying the spatial patterns of ecosystem services and their driving factors is crucial for strengthening ecological management and promoting sustainable environmental development. This research focuses on Beijing as the study area. The InVEST model was applied to analyze the spatial correlation， trade-offs， and synergies of habitat quality， carbon storage， water yield， and soil retention from 2000 to 2020. The analysis utilized methods such as spatial autocorrelation， cold/hot spot analysis， and bivariate spatial autocorrelation analysis. Additionally， the XGBoost-SHAP model was employed to identify the key factors affecting ecosystem services. The results showed that： ① The high-value areas of habitat quality were mainly concentrated in regions with higher terrain and less interference from human activities. Carbon storage exhibited a spatial distribution trend that was high in the northwest and low in the southeast. The high-value areas of water yield were concentrated in urban areas， while the high-value areas of soil conservation were primarily distributed in the southwest and were more scattered in the north. ② Global spatial autocorrelation analysis indicated that the global Moran's I indices for the four ecosystem services all passed the significance test and demonstrated significant high-value aggregation characteristics. ③ There was a significant synergistic relationship between habitat quality， carbon storage， and soil conservation. However， there was a trade-off between water yield and these factors. ④ The XGBoost regression model showed good prediction performance on both the training set and the test set， with the predictive performance on the training set being better than that on the test set. The SHAP model analysis indicated that elevation was the key driving factor affecting the four ecosystem services. Slope significantly affected habitat quality， carbon storage， and soil conservation. Population density mainly affected habitat quality and water yield， while annual precipitation had an important influence on water yield and soil conservation. The research results can provide scientific support for optimizing the spatial patterns of ecosystem services and formulating ecological protection strategies in Beijing.

Drawing on 2011-2023 panel data from multiple Chinese provinces, this study empirically tests the impact of digital inclusive finance (DIF) on the urban-rural income gap (URIG) through the lens of agricultural productive services (APS). The findings indicate that DIF well reduces URIG, though such effect varies across regions and DIF dimensions. Further analysis reveals that DIF narrows income gaps primarily by expanding rural credit access and improving agricultural productivity. Moreover, DIF exhibits notable spatial spillover effects, reducing income inequality within local areas alongside promoting positive externalities in neighboring regions. Consequently, it is essential to boost the localized implementation of high-quality DIF initiatives, leverage spatial spillover effects to foster coordinated regional development, actively bridge the urban-rural digital divide, and further integrate DIF into rural economic revitalization strategies.

Accurate functional annotation of viral proteins is essential for genomic surveillance, yet rapid viral evolution causes "functional drift" that challenges conventional sequence-only models. These models often lack interpretability and struggle with fragmented sequences from complex environmental samples such as wastewater. We developed ViralMultiNet, a structure-aware multimodal framework that integrates multi-scale k-mer encodings (4-7-mers) with functional semantic embeddings derived from UniProt annotations. Using a curated Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) dataset of 66,011 samples from wastewater metagenomics (NCBI SRA: SRX28474964), we implemented gated multimodal fusion and triple knowledge distillation to transfer structural insights from a teacher to a student model. Model performance was evaluated via 5-fold cross-validation and external validation on emerging variants. Training efficiency was optimized using Low-Rank Adaptation and Flash Attention. ViralMultiNet achieved robust classification performance with a macro F1 score of 0.921 ± 0.004, accuracy of 0.928 ± 0.003, and AUC of 0.983 in cross-validation. The distilled student model matched teacher performance within a negligible margin (<0.003 F1 difference) while reducing training time by 40.4% (from 94.3 to 56.2 minutes per epoch). Interpretability analysis revealed that model attention peaks consistently aligned with experimentally validated functional domains of the SARS-CoV-2 Spike protein, including the receptor-binding domain (residues 319-541), S1/S2 cleavage site (681-685), and fusion peptide (816-835). ViralMultiNet offers a scalable, interpretable solution for viral protein function prediction. Its ability to generalize across variants and map attention to critical biological regions supports deployment in wastewater-based early warning systems, enhancing global pandemic preparedness.

Accurate identification and classification of tomato leaf diseases play a pivotal role in increasing tomato yield. However, existing tomato leaf disease detection approaches face two major challenges, one is inadequate model ing of underlying higher-order disease correlations, and the other is vulnerability to environmental interferences including leaf occlusion and illumination variations. To address these challenges, we propose a novel model, i.e., HSMK-YOLO, an enhanced framework based on YOLOv11 with three key innovations: firstly, a dual-branch architecture combines a hypergraph convolutional network for capturing higher-order disease relationships with the Kolmogorov-Arnold Network (KAN) for more efficient feature extraction; secondly, convolutional blocks re designed by spatial-channel reconstruction employ spatial attention mechanisms to enhance cross-feature map fusion; finally, traditional multilayer perceptrons (MLPs) are replaced with a Multi-Self-supervised Equivariant Attention Mechanism (MultiSEAM) to boost leaf disease identification and classification accuracy. Experimental results demonstrate that HSMK-YOLO achieved a precision of 92.1% and an mAP50 of 94.7%, outperforming Hyper-YOLO, the benchmark model, by 5.4% and 2.5%, respectively. Systematic comparison and generalization tests were conducted to further verify the effectiveness and generalization capability of the proposed model. The results show that HSMK-YOLO was superior to state-of-the-art models including YOLOv8, YOLOv9, YOLOv10, RT-DETRv2, and YOLOv12 in multiple metrics, fully demonstrating that the HSMK-YOLO model has outstand ing detection performance and broad application prospects, capable of providing powerful technical support to tomato disease control.