Riparian ecosystems encompass a diverse suite of ecosystem types, including river banks, floodplains, and wetlands, that are characterized primarily by being ecotones, or transitional zones, between adjacent terrestrial and aquatic realms. In general, riparian ecosystems represent wetter, cooler, and more heterogeneous habitats than adjacent upland areas and consequently tend to support biologically distinctive, productive and diverse communities. Riparian ecosystems also provide a wealth of critical ecosystem functions and services of importance at local and catchment scales. In particular, riparian vegetation plays a key role in: regulating microclimates and water quality; preventing riverbank erosion and promoting landform stability; subsidizing aquatic and terrestrial food webs; and providing habitat for a wide range of aquatic, amphibious, and terrestrial organisms. Despite these values, riparian ecosystems are also some of the most altered, degraded and vulnerable ecosystems on Earth, due largely to their position in the landscape and hotspots of intensive human activity. Effective management and restoration of riparian ecosystems is increasingly urgent, for both the conservation of much terrestrial and aquatic biota and the wellbeing and livelihoods of people throughout the world. Management approaches range from protection of intact riparian ecosystems to active revegetation of degraded riparian zones. Significant knowledge gaps exist in many places, however, regarding relationships between riparian ecosystem structure and function and the responses of these to disturbances and management interventions.
Self-supervised Learning (SSL) has become a powerful paradigm for representation learning without manual annotations. However, most existing frameworks focus on global alignment and struggle to capture the hierarchical, multi-scale lesion patterns characteristic of plant disease imagery. To address this gap, we propose PSMamba, a progressive self-supervised framework that integrates the efficient sequence modelling of Vision Mamba (VM) with a dual-student hierarchical distillation strategy. Unlike conventional single teacher-student designs, PSMamba employs a shared global teacher and two specialised students: one processes mid-scale views to capture lesion distributions and vein structures, while the other focuses on local views to capture fine-grained cues such as texture irregularities and early-stage lesions. This multi-granular supervision facilitates the joint learning of contextual and detailed representations, with consistency losses ensuring coherent cross-scale alignment. Experiments on three benchmark datasets show that PSMamba consistently outperforms representative CNN-, Transformer-, SSL-, and Mamba-based baselines, delivering superior accuracy and robustness in both domain-shifted and fine-grained scenarios.
The literature has frequently studied the effect of privacy concerns, irritation, trust, incentives, enjoyment, innovativeness, usefulness, ease of use and subjective norms on mobile advertising value. However, the findings are inconsistent regarding the real effect of these antecedents, leaving practitioners uncertain about which antecedents to rely upon to design effective mobile advertisements triggering a problem for them. Using a meta-analytic approach, the current research synthesises 211 empirical studies to test the relationship between these antecedents and advertisement value, which subsequently influences attitudes, intentions, and actual behaviour of receiving mobile advertisements. In terms of findings, our study suggests that usefulness is the strongest driver of advertisement value, whereas irritation is the strongest mitigator, and subjective norms have minimal influence. Similar patterns were observed for the effect of the antecedents on attitudes and intentions. Findings also suggest that strength of these relationships varies depending on contexts and methodological approaches adopted in prior studies. Specifically, the effects of antecedents on the outcome variables are comparatively stronger in recent studies, published in high-quality journals, in developed countries, younger samples, experimental designs, smartphone contexts, customized ads, permission-based and location-aware ads, and in-app formats. Accordingly, this study resolves the inconsistent findings of previous research which makes a novel contribution to the mobile advertising literature. Regarding implications, we recommend mobile advertisers use precision targeting using customers’ geographical area and age group, prioritise customised content over generic campaigns, and adopt permission-based communication as a standard practice.
Graph Neural Networks (GNNs) often struggle with node classification under class imbalance because standard methods fail to adequately leverage graph topology. Specifically, majority nodes dominate message passing, causing the aggregated information from minority nodes to become statistically diluted, which masks relational signals associated with under-represented classes. Traditional techniques such as oversampling and cost-sensitive learning mitigate data skew but typically neglect structural dependencies, leading to under-representation of minority classes or overfitting to local graph substructures. We propose CL3E-GNN, a novel curriculum learning framework that addresses class imbalance through three synergistic phases: Engage, Enact, and Embed. The Engage phase employs multi-head attention to identify critical node and edge connections, generating enhanced feature representations via dual graph convolution network (GCN) and graph attention network (GAT) aggregation. The Enact phase refines minority class representations through three complementary mechanisms, each targeting a distinct source of imbalance: importance sampling for active feature learning, metapath propagation for knowledge transfer across heterogeneous relations, and anomaly detection for identifying structural irregularities. The Embed phase introduces a memory-augmented reinforcement learning mechanism that progressively refines node embeddings via reward-modulated updates. A dual-loss optimisation strategy combining cross-entropy with adaptive curriculum loss ensures gradual difficulty progression, offering a principled alternative to simple loss functions or data augmentation alone. Extensive evaluations on eight benchmark datasets, with results averaged over five independent runs, demonstrate consistent improvements over state-of-the-art methods. These results establish CL3E-GNN as a new benchmark for imbalanced node classification, effectively preserving topological integrity while mitigating class distribution skew. The implementation code is available at https://github.com/afofanah/CL3E-GNN.
Artificial intelligence (AI) is transforming service industries by enabling personalised experiences that enhance customer satisfaction and operational efficiency. Despite its growing significance, research on AI-powered personalisation remains fragmented, being dominant by technicality and with limited consensus on conceptualisation, theoretical frameworks, key concepts, and typologies. This paper addresses these shortcomings through a structured review of 201 journal articles published between 2006 and 2024 using the Theory-Context-Method-Characteristic (TCMC) framework. The review offers valuable theoretical contributions and practical implications. First, it bridges service and computer science literature through a unified analytic framework and develops an updated conceptualisation of AI-powered personalisation. Second, major research gaps are identified, including limited theory-driven inquiry, concentration on tourism and hospitality contexts, an overreliance on algorithm optimisation methods (i.e., algorithm-based experiments), and insufficient attention to personalisation features, consumer psychology, and non-customer stakeholders. Third, a forward-looking agenda is proposed, emphasising stronger theoretical grounding, broader cross-industry/ cross-country research, and deeper examination of emotion-aware and anthropomorphic AI. Finally, for practitioners, the findings underscore that responsible and high-impact AI-powered personalisation hinges on three pillars: transparent data governance, intuitive and low-friction AI touchpoints, and a workforce equipped to manage and complement advanced AI systems.
