Zhejiang Gongshang University

Parent-child argumentation is a unique form of communication, as it combines persuasion, emotional exchange, and instructional dynamics shaped by contextual factors and cultural norms. To fully understand how a parent balances dialectical reasonableness with rhetorical adaptability to resolve conflicts and foster cooperation with the child within specific context, this study investigates the argumentative style of parent-child interaction within the Chinese context, focusing on the interplay of cultural values, educational goals, and argumentative practices in family interaction. Within a corpus of 20 hours of recordings of parent-child conversations concerning educational topics from 5 Chinese families, a conversation between a mother and a son during homework was selected and qualitatively analyzed, based on the framework of Argumentative Style developed from the standard model of pragma-dialectics. The findings highlight the predominant use of an engaged argumentative style in the case, which features the parent’s radiating commitment in topical selection, communality in adaptation to audience demand and inclusiveness in presentational devices. An occasional shift to a detached style was also identified, particularly when the authority figure of the teacher was invoked. The subtle balance between nurturing parental involvement and reinforcing respect for established norms is reflective of the broad Chinese cultural values that parents bear significant responsibilities for children’s academic success, and act as guides and enforcers in family education. By investigating the roles different argumentative styles play in real-life parent-child interaction, this study provides implications for developing effective communication strategies in family education and highlights the significance of culturally informed argumentative practices.

The advent of the information age has led to the growing development of digital trade, which has a potential impact on the position of the service sector in global value chains (GVCs). In order to further clarify the mechanism of this impact, this paper selects a sample of 50 economies in the ADBMRIO database for the period of 2005-2020 and conducts a benchmark regression analysis with the GVC status index and digital trade as the explanatory variables and the core explanatory variables, respectively. The results of the analysis show that the relationship between digital trade and service industry GVC is “inverted U”, that is, in the early stage of the participation of digital services in global trade, the increase of digital trade has increased the status of service industry GVC, and with the deepening of the degree of digitization, the status of the service industry GVC declines with the increase of the level of digital trade. The following are a few examples of how digital trade affects services. Also for lower-middle-income, upper-middle-income, and high-income countries, the coefficients of the impact of digital trade on the global value chain position of the service industry are 0.152, 1.752, and -0.022, respectively, revealing that the relationship between digital trade and the service industry’s position in the full value chain varies with different income levels of the country.

Accurately extracting lesions from medical images is a fundamental but challenging problem in medical image analysis. In recent years, methods based on convolutional neural networks and Transformer have achieved great success in the medical image segmentation field. Combining the powerful perception of local information by CNNs and the efficient capture of global context by Transformer is crucial for medical image segmentation. However, the unique characteristics of many lesion tissues often lead to poor performance and most previous models failed to fully extract effective local and global features. Therefore, based on an encoder-decoder architecture, we propose a novel alternate encoder dual decoder CNN-Transformer network, AD2Former, with two attractive designs: 1) We propose alternating learning encoder can achieve real-time interaction between local and global information, allowing both to mutually guide learning. 2) We propose dual decoder architecture. The unique way of dual-branch independent decoding and fusion. To efficiently fuse different feature information from two sub-decoders during decoding, we introduce a channel attention module to reduce redundant feature information. Driven by these two designs, AD2Former demonstrates strong capture ability for target regions and fuzzy boundaries. Experiments on multi-organ segmentation and skin lesion segmentation datasets also demonstrate the effectiveness and superiority of AD2Former.

Government-led administrative division adjustment has occurred throughout the various stages of China’s social development and serves as a significant driving force for urbanization in the country. As representative of administrative division adjustment, the Revoke County to Urban District (RCUD) policy has played a crucial role in the urbanization process in China. The question must therefore be asked as to whether RCUD does indeed enhance urban land use efficiency (ULUE). This has significant implications for China’s new-type urbanization strategy and high-quality economic development. This paper constructs a research framework on the relationship between RCUD and ULUE, and empirically examines the effect of RCUD on ULUE using data from Chinese prefecture-level and above cities. The results show that RCUD significantly promotes improvements to ULUE, and this conclusion still holds after a series of robustness and endogeneity tests. The policy effect of RCUD exhibits heterogeneity, not only in terms of economic development but also in terms of city size and in terms of RCUD reform experience. The effect improvements to RCUD have on ULUE is more pronounced in economically developed cities. Furthermore, RCUD significantly increases ULUE only in large cities and above; the impact of RCUD is greater in cities with prior experience. Mechanism tests reveal that promoting elements agglomeration and optimizing industrial structure are important channels by which RCUD can enhance ULUE; however, the disorderly expansion of boundaries triggered by RCUD has adverse effects on ULUE. The research findings provide theoretical support and reference for optimizing further RCUD reform and enhancing ULUE.

Global food loss and waste continues to increase despite efforts to reduce it. Food waste causes a disproportionally large carbon footprint and resource burdens, which require urgent action to transition away from a disposal-dominated linear system to a circular bioeconomy of recovery and reuse of valuable resources. Here, using data from field-based studies conducted under diverse conditions worldwide, we found collective evidence that composting, anaerobic digestion and repurposing food waste to animal feed (re-feed) result in emission reductions of about 1 tCO2e t⁻¹ food waste recycled compared with landfill disposal. Emission mitigation capacity resulting from no landfill disposal in the United States, the European Union and China would average 39, 20 and 115 MtCO2e, which could offset 10%, 5% and 17% of the emissions from these large agricultural systems, respectively. In addition, re-feed could spare enormous amounts of land, water, agricultural fuel and fertilizer use. Our findings provide a benchmark for countries developing food waste management strategies for a circular agrifood system.

Studies suggest that family motivation can increase employees’ performance because it integrates work and family roles. However, the literature overlooks the possibility that role integration between work and family may also trigger work-family conflict, which can decrease performance. To address this oversight, this study explores how and when family motivation influences employees’ in-role (i.e., task performance) and extra-role performance (i.e., helping behavior) through work-family conflict. Two field studies were conducted in China to test our hypotheses. Study 1 surveyed 362 employees and found that family motivation exerted an indirect adverse effect on both task performance and helping behavior through its impact on work-family conflict. Study 2 surveyed a sample of 481 employees and found that family motivation could stimulate work-family conflict via role integration, and that this mediated effect was stronger for employees with a high level of segmentation preference. This study advances the research on family motivation, and offers insights into ways to effectively manage work-family relationships.

With the advancement of digital technologies, tourism live streaming (TLS) has rapidly gained global popularity due to its real-time and interactive features, showcasing significant marketing potential. However, viewer retention remains a major challenge and a bottleneck for TLS development. This study first defined the concept and dimensions of alternative attractiveness in TLS through qualitative interviews. It then constructs an analytical framework based on the Push–Pull–Mooring (PPM) theory. It empirically tests how psychological contract breach, viewer-live streamer social distance, and alternative attractiveness influence viewers’ non-continuous following intention (NCFI) in TLS. The findings reveal that these factors significantly impact NCFI. Customer complaining behavior mediates the relationship between psychological contract breach and NCFI, and perceived controllability positively moderates this relationship. This study provides a new theoretical perspective on understanding viewer attrition mechanisms and offers practical suggestions for TLS platforms and streamers to enhance viewer retention.

Anemia is globally linked to dietary iron deficiency, potentially concerned by a shift from meat-based diets to plant-based ones with less bioavailable non-heme iron. This study compared the iron bioavailability of two commercial plant-based burgers (PBB1 and PBB2) with that of an animal-based burger (ABB). PBB1 and PBB2 contain 2.37 mg and 2.45 mg, respectively, while ABB contained 1.6 mg of iron per 100 g. The iron bioavailability (ng ferritin/mg protein) of PBB2 (5.98 ± 0.41) and PBB1 (4.70 ± 0.33) was higher than ABB (4.05 ± 0.29) as determined using a Caco-2 cell model. The main inhibitors and enhancers of iron bioavailability were also investigated. Phenolic compounds were found to increase iron bioavailability in the PBBs, suggesting they may not always act as antinutritional factors. Phytic acid content had no significant impact on iron bioavailability. There was a positive correlation between the antioxidant properties of the digested burgers and iron bioavailability. These findings suggest that PBBs can match or exceed the iron bioavailability of ABB, offering potential solutions for global nutritional challenges.

Deposit insurance pricing is crucial for the successful implementation of a deposit insurance system. Based on the Merton deposit insurance option pricing framework, we present a deposit insurance pricing model that simultaneously considers multiple factors. The proposed model is more general than existing models, as it can degenerate into simpler models, such as those involving only a single factor. Given that the priorities for the repayment of deposits within and beyond the insurance limit should differ, we incorporate the insurance limit factor into the pricing model from the perspective of the maximum underwriting amount. This approach aligns with the original intent of setting an insurance limit in deposit insurance systems. The illustration using selected commercial banks in China shows the advantages of the proposed model.

Timely identification of Alternaria alternata infection in postharvest green peppers is crucial before overt symptoms. This study evaluated peppers’ appearance, chlorophyll, chlorophyll fluorescence parameters, and membrane leakage after A. alternata infection. Fatty acids of thylakoid membrane in peppers were determined by gas chromatography-mass spectrometry (GC–MS). Diseased fruit showed early declines in appearance, chlorophyll, fluorescence parameters, and membrane integrity, with minimum fluorescence (F0), electrical conductivity, and the absorbance of nucleic acid at 260-nm wavelength (OD260) distinguishing them from the controls within 1 d. These changes preceded sensory quality and other chlorophyll metrics. Diseased areas exhibited significant differences than healthy areas. Total fatty acid contents in the controls increased, while those in diseased fruit decreased. Linoleic acid had a significant correlation with membrane leakage, fluorescence, and chlorophyll, respectively. The degradation of unsaturated fatty acids led to significant changes in chlorophyll fluorescence parameter and membrane leakage. F0, electrical conductivity, and OD260 were all suitable for early identification of A. alternata infection in postharvest peppers.

With the increasing complexity of user-item interactions on the Internet, it is important to profile users and model their preferences in recommender systems. Traditional methods, including metric learning, rely on historical user-item interactions to model preferences but struggle in sparse data scenarios. While item tags offer valuable auxiliary information to enhance representations, their shared nature across items makes it challenging to effectively profile users with tags, which requires preserving user personalization through high-quality tag representations. Moreover, traditional optimization for user/item representations always takes place in Euclidean space, where the unconstrained nature of embedding norms tends to lean toward trivial solutions. This may bias the system towards common or popular preferences, thus suppressing the variety in tag-aware user profiles. To this end, we propose to profile users with tag-enhanced spherical metric learning for recommendation, named UTRec. Specifically, we propose an adaptive tag selection mechanism to ensure the quality of tag representations and learn tag-enhanced representations of users/items, thereby effectively profiling users. Additionally, we introduce a spherical optimization strategy for tag-enhanced recommendations to alleviate the limitations imposed by lazy learning and traditional optimization, ensuring the accuracy and diversity of user and item representations within the spherical space. Numerous experiments have been conducted on four real-world datasets, where our proposed tag-enhanced UTRec framework can bring consistent performance gains and achieve a 13.67% improvement regarding both Recall and NDCG metrics.

Autonomous delivery vehicles (ADVs) that provide contactless services have attracted much academic and practical attention in China in recent years. Despite this, there is a lack of in-depth research on what motivates customers to embrace ADVs. The study integrates the theory of planned behavior (TPB) and normative activation model (NAM) and explores how environmental factors, situational factors, and individual factors affect original TPB constructs and ultimately consumers’ intention to use ADVs. Structural equation modeling was performed on survey data of 561 Chinese consumers through an online sampling platform. The results show that among the factors affecting consumer intention, word-of-mouth recommendations have the greatest impact, followed by perceived enjoyment, COVID-19 risk, ascription of responsibility, subjective norm, attitude, and perceived behavioral control. The results not only make important theoretical contributions to the technology acceptance fields but also provide helpful references to logistics enterprises, ADVs technology providers, and policymakers.

The fusion of hyperspectral image (HSI) and light detection and ranging (LiDAR) data offers a powerful approach for land cover classification; however, challenges remain in effectively integrating their complementary information. Existing methods often overlook the importance of spatial information and fail to fully exploit the synergy between HSI and LiDAR data. To address these limitations, this paper proposes M²SSCENet, a multi-branch multi-scale joint learning and spatial-spectral cross-enhancement network. M²SSCENet employs a three-branch architecture to extract HSI spectral features, HSI spatial features, and LiDAR features, respectively. For cross-modal fusion, the network proposes two novel modules: the cross-modality bilateral attention feature fusion module enhances the interaction between HSI spectral features and LiDAR features, while the spatial attention-guided cross-modality fusion module dynamically adjusts spatial attention to capture key elevation information. Additionally, a pixel distance-based proximal feature selection module is proposed to enhance spatial feature representation by emphasizing neighboring pixels with higher contributions. Experimental results on the Trento and Houston2013 datasets demonstrate the superiority of M²SSCENet, achieving OA of 98.44% and 94.33%, respectively. Compared with suboptimal methods on each dataset, M²SSCENet improves classification accuracy by 0.27% on the Trento dataset and by 2.03% on the Houston2013 dataset. Notably, for categories with similar spectral distributions but significant elevation differences, such as “Highway” and “Parking Lot 1,” the proposed method achieves accuracy improvements of 2.18% and 4.75%, respectively. These results highlight the effectiveness of M²SSCENet in leveraging the complementary strengths of HSI and LiDAR data for improved land cover classification.

Multi-object tracking requires accurately identifying and tracking multiple targets over long periods. However, tracking performance is highly susceptible to various factors, such as target deformation, occlusion, etc. Meanwhile, most existing MOT models perform simple aggregation and classification of target features, ignoring the inherent differences and connections between detection and re-identification. This often leads to frequent identity switches. To address the above issues, we propose our tracker IFMOT, a simple and efficient network that combines an interactive perception network with feature optimization. Specifically, we propose an interactive perception network with a multi-head cross-attention mechanism design to alleviate feature conflicts. And then, we introduce a feature optimization module that refines the target representation to improve the extraction capability of feature embeddings. Furthermore, a feature integration similarity matrix is used to comprehensively assess the similarity between objects and handle unreliable similarity matching. Experiments on the MOT16, MOT17, MOT20 and Dancetrack datasets show that the proposed method achieves a higher accuracy while keeping the tracking speed, in contrast to other state-of-the-art trackers.

Against the backdrop of rapid mobile internet evolution, managing the virtual supply chain of mobile applications (apps) is increasingly critical. This study compares the impacts of two operational modes—traditional channel distribution and the rapidly emerging traffic purchase mode, fueled by the prevalence of information flow advertising—on mobile app market performance. It considers two types of mobile apps: one-time payment and continuous payment models. By constructing and deriving mathematical models, this research explores the game equilibrium between mobile app developers, channel distributors, and traffic purchase platforms under both channel distribution and traffic purchase modes. We also analyze the performance of different virtual supply chain operation modes for varying types of mobile apps. Furthermore, this study investigates the role of app quality and user retention rate in supply chain decision-making. The results indicate that app pricing is lower under the channel distribution mode compared to the traffic purchase mode. Developers should opt for the traffic purchase mode when consumer retention rates or quality preferences are low; otherwise, the channel distribution mode is preferable. By considering different operational modes' game structures, this study provides valuable managerial insights for practitioners in the mobile app field.