Search Results (422)

Recently, recommendation systems have actively integrated Transformers to capture real-time context. However, these systems often suffer from generalization imbalance, where predictions are biased toward popular (head) items due to the sparsity and volatility inherent in session-based data. To address this challenge, this paper proposes MoE-SLMRec, a Mixture-of-Experts (MoE)-based recommendation model that selects expert networks based on session-level contextual information. The proposed model extracts a session latent representation, h, through a session-aware controller and forms balanced predictive characteristics across the entire data distribution via dynamic routing. Experimental results demonstrate that MoE-SLMRec significantly outperforms the baseline SLMRec, improving accuracy by 1.51 percentage points (from 18.76% to 20.27%). Furthermore, the model achieved state-of-the-art performance in Recall@20 (0.8358) and MRR@20 (0.3455), validating simultaneous improvements in both retrieval capability and ranking quality. Notably, the model effectively stabilized the performance for head items while coordinating the generalization trade-off between head and tail segments. By ensuring a favorable capacity–cost trade-off while maintaining robust performance, this study presents a promising alternative under session-based recommendation settings, facilitating scalable deployment in real-time recommendation services. Full article

Earthen materials are attractive sustainable building solutions due to their low embodied energy and ecological benefits. However, their inherent weaknesses, such as low strength and poor durability, severely restrict modern engineering applications. Traditional physical or chemical modification methods struggle to balance significant improvement in mechanical performance with the preservation of their core sustainable attributes. To overcome this long-standing challenge, this study proposes a paradigm-shifting solution: a prefabricated monolithic lattice–earth composite wall structure. This system abandons the single-material-centered modification approach. Instead, through macroscopic system-level composite design, reinforced concrete lattices and earthen blocks are prefabricated into integral wall panels in a factory. These panels then work collaboratively with the peripheral frame through reliable integral connections. Via quasi-static tests and theoretical analysis on four scaled wall specimens with different design parameters, this study systematically reveals the working mechanism and performance regulation principles of this composite system. The core findings indicate: (1) The system achieves multiple seismic defense lines and a controllable energy dissipation path through a sequential damage mechanism: “earthen material cracking and friction → lattice yielding and energy dissipation → final defense by the frame.” (2) The ratio of the equivalent lateral stiffness of the prefabricated wall panel to the stiffness of the outer frame is a key dimensionless design parameter controlling the failure mode (ductile shear or brittle bending), and the lattice configuration is an effective means to adjust this parameter. (3) Based on tests and an equivalent stiffness model, quantitative design guidelines are proposed, focusing on optimizing lattice density (recommended: 3–4 lattice columns), limiting the aspect ratio (preferably ≤1.5), and ensuring “strong connections.” This study demonstrates that the system, without sacrificing the intrinsic sustainable advantages of earthen materials, successfully endows them with high performance, meeting modern seismic code requirements and potential for prefabricated construction through system integration innovation. It provides a new path with theoretical foundation and practical feasibility to resolve the core contradiction in the modernization of traditional earthen buildings—the incompatibility between ecological attributes and engineering performance. This lays an important foundation for developing next-generation high-performance green building structural systems. Full article

Personalized recommender systems play an essential role in enhancing user experience by accurately predicting user preferences. Previous approaches mainly focus on modeling long-term preferences or capturing short-term dynamics through sequential patterns, while few achieve an effective balance between the two. This study proposes Rec-SSP, a novel review-aware recommendation model that integrates long-term and short-term preferences through a gated fusion mechanism. Long-term preferences are extracted from aggregated user reviews, whereas short-term preferences are modeled by identifying sequential patterns from recent interactions at both the review and category levels. This multilevel design captures fine-grained opinions across items, ensuring a more accurate understanding of the evolving user intent. This study conducted various experiments on real-world datasets, showing that Rec-SSP outperforms baseline models. These findings demonstrate that balancing long-term and short-term preferences with multilevel sequence modeling can significantly improve recommendation accuracy across diverse domains. Full article

This paper proposes a multi-behavior and sequence-aware recommendation method that effectively integrates diverse user–item interaction behaviors and their sequential dependencies to enhance recommendation accuracy. Unlike existing studies that treat different user–item interactions independently, our approach integrates diverse behaviors and their natural sequential dependencies to better capture user preferences and alleviate data sparsity caused by single-behavior modeling. Different from the traditional single-behavior models, our approach constructs a multi-behavior heterogeneous graph and defines multiple meta-path patterns to capture implicit relationships between users and items. By generating subgraph instances, we extract fine-grained interaction patterns and employ a LightGCN with residual connections to learn user representations under different behavioral sequences. Furthermore, an attention mechanism is introduced to fuse features across subgraphs, enabling more expressive preference modeling. Experimental results on two real-world datasets, Taobao and Tmall, demonstrate that our method outperforms state-of-the-art single- and multi-behavior recommendation models, achieving up to 10.0% and 11.1% improvements in HR@10 and NDCG@10 on Taobao and 9.0% and 10.6% on Tmall, respectively. These results confirm the effectiveness of leveraging both multi-behavior information and sequence dependencies in capturing deeper user preferences for more accurate recommendations. Full article

The rapid development and deep integration of digital technology into cultural heritage have created new experiential paradigms for tourists. However, to transform from technological application to behavioral retention, the internal mechanisms through which digital experiences are internalized into stable, sustained behavioral intentions must be elucidated. The influence of perceived value on tourists’ long-term behavioral intentions via place attachment remains largely unexplored. Using the value adoption model (VAM), this study constructs a sequential mediation model of “digital experience–perceived value–place attachment–sustained behavioral intentions” and employs structural equation modeling to examine cross-sectional survey responses from 618 tourists visiting Shandong Museum, China. Findings reveal that the functional dimensions of interactive experience and perceived ease of use significantly enhance perceived value, whereas the sensory dimensions of immersive and hedonic experiences have no significant impact on perceived value—possibly because tourists in cultural heritage contexts prioritize knowledge acquisition over sensory stimulation. Perceived value significantly and positively predicts place attachment and sustained behavioral intentions, and place attachment strongly predicts sustained behavioral intentions (including word-of-mouth recommendation, revisit intention, and sharing). This study extends the VAM to offline cultural heritage digital experience contexts, demonstrates that functional utility is more critical than sensory stimulation in driving value perception, and validates the value attachment–behavior transformation pathway, providing theoretical foundations and practical implications for cultural heritage digitalization management. Full article

This paper presents an efficient parameter optimization approach to the plastic injection molding process to achieve high productivity. In collaboration with a company specializing in plastic injection-mold-based production, real process data was collected and used in this research. The result is an integrated framework, combining a genetic algorithm (GA) with a CatBoost-based surrogate model for multi-objective optimization of the injection molding machine parameters. The aim of the optimization is to minimize the cycle time and cycle energy while maintaining the product quality. Ten process parameters were optimized, which are machine-specific. An evolutionary optimization using the NSGA-II algorithm is used to generate the recommended parameter set. The proposed GA-surrogate hybrid approach produces the optimal set of parameters that reduced the cycle time by 4.5%, for this specific product, while maintaining product quality. Cycle energy was evaluated on an hourly basis; its variation across candidate solutions was limited, but it was retained as an optimization objective to support energy-based process optimization. A total of 95% of the generated solutions satisfied industrial quality constraints, demonstrating the robustness of the proposed optimization framework. While classical Design of Experiment (DOE) approaches require sequential physical trials, the proposed GA-surrogate framework achieves convergence in computational iterations, which significantly reduces machine usage for optimization. This approach demonstrates a practical way to automate data-driven process optimization in an injection mold machine for an industrial application, and it can be extended to other manufacturing systems that require adaptive control parameters. Full article

Background/Objectives: The treatment of dual disorders (DDs) must be comprehensive and multidisciplinary. Evidence supports the effectiveness of psychotherapy in treating DDs. The second part of this consensus synthesizes the available evidence on psychological treatment for specific DDs. Methods: Two consensus methods were sequentially implemented: the nominal group technique and the Delphi method. Results: This consensus review encompassed a compilation of recommendations for the psychological treatment of neurodevelopmental, anxiety, post-traumatic stress, somatic symptom, eating, and personality disorders. Finally, recommendations for the future research agenda on the psychological treatment of DD were included. Conclusions: (1) Psychological treatment, particularly integrated treatment, is effective. (2) In the case of dual autism, interventions for substance use disorders should be adapted to this population’s characteristics. (3) More research is needed on dual social anxiety, panic, generalized anxiety, somatic symptom, and eating disorders, for which Cognitive Behavioral Therapy (CBT) is the most commonly used treatment. (4) For dual attention deficit hyperactivity disorder, multicomponent treatment is recommended (psychoeducation, CBT, and peer or family support). (5) For dual anxiety disorders, CBT is the first-line treatment. (6) For dual post-traumatic stress disorder, CBT (cognitive processing therapy and prolonged exposure therapy), acceptance and commitment therapy, stress inoculation training, and Eye Movement Desensitization and Reprocessing (EMDR) are effective. (7) For dual personality disorders, evidence is scarce. (8) For borderline personality disorder, dialectical behavior therapy, dynamic deconstructive psychotherapy, and dual-focus schema therapy show promise. (9) For antisocial personality disorder, CBT, contingency management, and counseling on impulsive lifestyles may be useful. (10) Much more evidence is needed from studies that overcome the methodological limitations of existing ones. Full article

This study investigates the challenge of affordable housing in Riyadh, a city undergoing rapid transformation aligned with Saudi Arabia’s Vision 2030. It aims to bridge the structural gap in the housing market by developing a comprehensive analytical framework that measures housing suitability for emerging middle-income families, linking it to economic, spatial, and behavioral dimensions. The research employs a sequential mixed-methods design. The first phase involved a Multi-Criteria Decision Analysis (MCDA) of 106 residential neighborhoods, constructing a Housing Suitability Index (HSI) based on financing cost (≤SAR 880,000), quality of urban life, and geographical accessibility. The second phase utilized focus groups with 16 participants from real estate developers and new families to explore behavioral drivers and subjective trade-offs. Quantitative results identified “convenience clusters” primarily in the city’s southeastern and southwestern sectors, offering an optimal balance between price and accessibility. Qualitative analysis revealed a significant trust gap and a misalignment of priorities: new families are increasingly willing to sacrifice unit size for central location and construction quality, a preference that conflicts with developers’ strategies focused on luxury units or peripheral projects for higher margins. The study concludes that achieving the 70% homeownership target requires a hybrid policy model, combining supply-side stimuli (e.g., subsidized land) with demand-side management (e.g., progressive mortgages). It recommends integrating the HSI into urban planning to direct investment towards logistically connected areas, fostering sustainable communities. Full article

Polar sea ice is undergoing rapid change, with recent record-low extents in both hemispheres, raising the demand for skillful predictions from days to seasons for navigation, ecosystem management, and climate risk assessment. Accurate sea ice prediction is essential for understanding coupled climate processes, supporting safe polar operations, and informing adaptation strategies. Physics-based numerical models remain the backbone of operational forecasting, but their skill is limited by uncertainties in coupled ocean–ice–atmosphere processes, parameterizations, and sparse observations, especially in the marginal ice zone and during melt seasons. Statistical and empirical models can provide useful baselines for low-dimensional indices or short lead times, yet they often struggle to represent high-dimensional, nonlinear interactions and regime shifts. This review synthesizes recent progress of DL for key sea ice prediction targets, including sea ice concentration/extent, thickness, and motion, and organizes methods into (i) sequential architectures (e.g., LSTM/GRU and temporal Transformers) for temporal dependencies, (ii) image-to-image and vision models (e.g., CNN/U-Net, vision Transformers, and diffusion or GAN-based generators) for spatial structures and downscaling, and (iii) spatiotemporal fusion frameworks that jointly model space–time dynamics. We further summarize hybrid strategies that integrate DL with numerical models through post-processing, emulation, and data assimilation, as well as physics-informed learning that embeds conservation laws or dynamical constraints. Despite rapid advances, challenges remain in generalization under non-stationary climate conditions, dataset shift, and physical consistency (e.g., mass/energy conservation), interpretability, and fair evaluation across regions and lead times. We conclude with practical recommendations for future research, including standardized benchmarks, uncertainty-aware probabilistic forecasting, physics-guided training and neural operators for long-range dynamics, and foundation models that leverage self-supervised pretraining on large-scale Earth observation archives. Full article

Modular construction refers to the use of factory prefabricated integrated module units. The modular steel construction unit SHS (Square Hollow Section) group column is a structure composed of four independent steel column units. Due to its compositional characteristics with voids, the fire resistance performance differs from ordinary steel columns, necessitating specific study. This paper employed a sequentially coupled thermal–mechanical analysis to investigate this. The effectiveness of the simulation model was first validated by comparing the simulated time–temperature curves and fire resistance limits with experimental results. A parametric analysis was then conducted to evaluate the influence of various factors, including the load ratio, cavity spacing, insulation type, gypsum board thickness, slenderness ratio, steel yield strength, and inner panel type, on the fire resistance limit. The results show that when the gypsum board thickness increased from 10 mm to 30 mm, the fire resistance limit correspondingly increased by 126%, 120%, 130%, and 130% for load ratios of 0.4, 0.5, 0.6, and 0.7, respectively. When the steel yield strength increased from 235 MPa to 690 MPa, the fire resistance limit increased by 20%, 21%, 24%, and 43% for load ratios ranging from 0.4 to 0.7. For inner panels of Glass Fiber, Rock Wool, Mineral Wool, and Plasterboard, the corresponding fire resistance limit ratios for load ratios of 0.4 to 0.7 were 1:1.13:1.24:1.45, 1:1.14:1.23:1.46, 1:1.11:1.2:1.42, and 1:1.08:1.18:1.41, respectively. It can be found that the best way to increase the fire resistance of the modular column is to increase the thickness of the gypsum board. A simplified calculation formula for the fire resistance limit of SHS group columns was derived through regression analysis, and recommendations for fire protection design were proposed, providing valuable insights for the future design and application of SHS group columns in steel modular construction. Full article

We present a combined ML/NLP (Machine Learning, Natural Language Processing) pipeline for protecting cloud-based APIs (Application Programming Interfaces), which works both at the level of individual HTTP (Hypertext Transfer Protocol) requests and at the access log file reading mode, linking explicitly technical anomalies with business risks. The system processes each event/access log through parallel numerical and textual branches: a set of anomaly detectors trained on traffic engineering characteristics and a hybrid NLP stack that combines rules, TF-IDF (Term Frequency-Inverse Document Frequency), and character-level models trained on enriched security datasets. Their results are integrated using a risk-aware policy that takes into account endpoint type, data sensitivity, exposure, and authentication status, and creates a discrete risk level with human-readable explanations and recommended SOC (Security Operations Center) actions. We implement this design as a containerized microservice pipeline (input, preprocessing, ML, NLP, merging, alerting, and retraining services), orchestrated using Docker Compose and instrumented using OpenSearch Dashboards. Experiments with OWASP-like (Open Worldwide Application Security Project) attack scenarios show a high detection rate for injections, SSRF (Server-Side Request Forgery), Data Exposure, and Business Logic Abuse, while the processing time for each request remains within real-time limits even in sequential testing mode. Thus, the pipeline bridges the gap between ML/NLP research for security and practical API protection channels that can evolve over time through feedback and retraining. Full article

The study explores how consumer buying behavior is influenced by artificial intelligence (AI) personalization, with a specific focus on responsible and sustainability-aligned digital marketing. Using an explanatory sequential mixed-methods design, the study analyzes a quantitative survey and qualitative interviews. Results show that purchase behavior is strongly affected by exposure to AI messages—especially when recommendations are relevant, timely, and emotionally appealing—and by trust in AI, while perceived lack of trust inhibits purchasing. Qualitative findings underscore affective responses alongside ethical concerns, perceived transparency, and perceived control over data. Overall, the study shows that effective personalization depends not only on algorithmic sophistication but also on users’ sense of relevance and autonomy and on ethical data governance. The conclusions highlight sustainability-consistent implications for marketers: increase data transparency, segment customers by privacy sensitivity, and adopt accountable, consent-based personalization to build durable trust and loyalty. Future research should examine longitudinal effects and cultural differences, acknowledging limits of small purposive qualitative samples for generalization and exploring how consumer trust, ethical perceptions, and responses to AI personalization evolve over time. Full article

Large language models (LLMs) demonstrate outstanding capability in understanding natural language and show great potential in open-domain travel planning. However, when confronted with multi-constraint itineraries, personalized recommendations, and scenarios requiring rigorous external information validation, pure LLM-based approaches lack rigorous planning ability and fine-grained personalization. To address these gaps, we propose the Symbolic LoRA Travel Planner (SLTP) framework—an agent architecture that combines a two-stage symbol-rule LoRA fine-tuning pipeline with a user multi-option heuristic tree search (MHTS) planner. SLTP decomposes the entire process of transforming natural language into executable code into two specialized, sequential LoRA experts: the first maps natural-language queries to symbolic constraints with high fidelity; the second compiles symbolic constraints into executable Python planning code. After reflective verification, the generated code serves as constraints and heuristic rules for an MHTS planner that preserves diversified top-K candidate itineraries and uses pruning plus heuristic strategies to maintain search-time performance. To overcome the scarcity of high-quality intermediate symbolic data, we adopt a teacher–student distillation approach: a strong teacher model generates high-fidelity symbolic constraints and executable code, which we use as hard targets to distill knowledge into an 8B-parameter Qwen3-8B student model via two-stage LoRA. On the ChinaTravel benchmark, SLTP using an 8B student achieves performance comparable to or surpassing that of other methods built on DeepSeek-V3 or GPT-4o as a backbone. Full article

Background/Objectives: The treatment of dual disorders (DDs) must be comprehensive and multidisciplinary. There is evidence supporting the effectiveness of psychotherapy in their treatment. However, clinical guidelines, consensus statements, and reviews on the treatment of DDs typically devote considerably less space to psychological therapy than to pharmacological therapy. Therefore, this work aimed to synthesize the available evidence, recommendations, and clinical experience on the psychological treatment of DDs to reach a consensus. Methods: Two consensus methods were sequentially implemented: the nominal group technique and the Delphi method. Results: The first part of this consensus review encompassed a compilation of general recommendations for the psychological treatment of DDs, evidence on the efficacy of the most frequently used therapies, and recommendations for the psychological treatment of severe dual mental disorders. These disorders include schizophrenia and other psychotic disorders, bipolar disorders, depressive disorders, and obsessive compulsive disorders. Conclusions: (1) Psychological treatment is effective; (2) integrated psychological treatment is more effective; (3) motivational interviewing, cognitive behavioral therapy, and relapse prevention are the psychological interventions with the most supporting evidence; (4) the best alternative is multicomponent strategies; (5) the most frequently studied severe mental disorders are schizophrenia and depression; (6) for dual schizophrenia, motivational interviewing and integrated cognitive behavioral therapy combined with other components are recommended; (7) for dual depression, cognitive behavioral therapy with relapse prevention or motivational interviewing is recommended; (8) for dual bipolar disorder, group therapies with psychoeducation or relapse prevention and inclusion of the family, contingency management, and family intervention are recommended; (9) more empirical evidence is needed, especially for obsessive compulsive and schizoaffective disorders; and (10) more randomized clinical trials are needed to improve current methodological limitations. Full article

Chronic obstructive pulmonary disease (COPD) is a major health burden worldwide and in Taiwan, ranking as the third leading cause of death globally, and its prevalence in Taiwan continues to rise. Readmission within 14 days is a key indicator of disease instability and care efficiency, driven jointly by patient-level physiological vulnerability (such as reduced lung function and multiple comorbidities) and healthcare system-level deficiencies in transitional care. To mitigate the growing burden and improve quality of care, it is urgently necessary to develop an AI-based prediction model for 14-day readmission. Such a model could enable early identification of high-risk patients and trigger multidisciplinary interventions, such as pulmonary rehabilitation and remote monitoring, to effectively reduce avoidable early readmissions. However, medical data are commonly characterized by severe class imbalance, which limits the ability of conventional machine learning methods to identify minority-class cases. In this study, we used real-world clinical data from multiple hospitals in Kaohsiung City to construct a prediction framework that integrates data generation and ensemble learning to forecast readmission risk among patients with chronic obstructive pulmonary disease (COPD). CTGAN and kernel density estimation (KDE) were employed to augment the minority class, and the impact of these two generation approaches on model performance was compared across different augmentation ratios. We adopted a stacking architecture composed of six base models as the core framework and conducted systematic comparisons against the baseline models XGBoost, AdaBoost, Random Forest, and LightGBM across multiple recall thresholds, different feature configurations, and alternative data generation strategies. Overall, the results show that, under high-recall targets, KDE combined with stacking achieves the most stable and superior overall performance relative to the baseline models. We further performed ablation experiments by sequentially removing each base model to evaluate and analyze its contribution. The results indicate that removing KNN yields the greatest negative impact on the stacking classifier, particularly under high-recall settings where the declines in precision and F1-score are most pronounced, suggesting that KNN is most sensitive to the distributional changes introduced by KDE-generated data. This configuration simultaneously improves precision, F1-score, and specificity, and is therefore adopted as the final recommended model setting in this study. Full article