Search Results (743)

With the widespread adoption of intelligent terminals, edge devices, and distributed information systems in the financial domain, financial security sensing data exhibit multisource heterogeneity, dynamic temporal patterns, and high privacy sensitivity. Traditional centralized anomaly detection methods are no longer able to simultaneously satisfy the requirements of cross-institutional or cross-node collaborative modeling, client data privacy protection, and robust monitoring of transaction and system anomalies. To address this challenge, a data-local federated deep anomaly detection framework has been proposed for distributed financial security sensing systems. Initially, a local deep financial security sensing representation module is constructed to perform temporal encoding and attention-based modeling on multisource financial signals, including terminal operation status, network transaction communication, backend server operation, identity authentication, and anomaly alerts, thereby extracting representations relevant to anomalous behaviors. Subsequently, a data-local federated optimization and personalized aggregation mechanism is developed to enable cross-node knowledge sharing without transmitting raw transaction or client data, while local personalized detection heads are employed to adapt to non-independent and identically distributed (non-IID) financial institution data. Furthermore, an adversarially robust security detection and trust-aware aggregation strategy is introduced to enhance model stability under input noise, feature masking, anomaly camouflage, and potential malicious client updates. Experimental results demonstrate that the proposed method achieves an Accuracy of $92.37\%$, a Precision of $89.41\%$, a Recall of $88.26\%$, an F1-score of $88.83\%$, an AUC of $93.06\%$, and a PR-AUC of $89.15\%$ in the primary financial anomaly detection task, significantly outperforming baseline methods such as Isolation Forest, Autoencoder, LSTM, Transformer, FedAvg, FedProx, SCAFFOLD, and MOON. In robustness experiments, the method attains F1-scores of $87.95\%$, $86.42\%$, $86.88\%$, $84.57\%$, $86.73\%$, and $83.91\%$ under Gaussian noise, feature masking, temporal shift, adversarial perturbation, and $20\%$ and $30\%$ malicious client scenarios, respectively. Ablation studies further confirm the effectiveness of local representation learning, personalized federated optimization, adversarial training, and trust-aware aggregation mechanisms. Overall, the proposed approach provides an efficient intelligent anomaly detection solution for financial AI security monitoring scenarios characterized by data localization requirements, node heterogeneity, and attack perturbations. Full article

Event extraction is a core NLP task that aims to identify triggers and arguments in unstructured text. In the financial domain, dense events, overlapping arguments, and ambiguous semantics pose significant challenges. This paper proposes CADAEE, a joint extraction framework that integrates curriculum adversarial learning and an enhanced adaptive layer. Curriculum adversarial learning dynamically adjusts training difficulty, thereby improving robustness and generalization on complex samples. The enhanced adaptive layer introduces learnable role-bias embeddings to model semantic dependencies between triggers and arguments, while a multi-head attention mechanism captures diverse feature interactions. Extensive experiments on the FewFC and DuEE-Fin datasets demonstrate the superiority of CADAEE. The model achieves highly competitive F1-scores in both trigger and argument classification, reaching 80.1% and 73.5% on FewFC, and 88.8% and 71.8% on DuEE-Fin, respectively. Ablation studies validate the synergistic contributions of the proposed modules. These results demonstrate that CADAEE provides robust and accurate extraction in complex, overlapping event scenarios, highlighting the value of combining curriculum learning with adaptive, role-aware enhancements for financial event extraction. Full article

Background/Objectives: In ketogenic therapy for children with refractory epilepsy—a special patient group—the quality of life of primary caregivers is often overlooked. This study aimed to explore the current state of primary caregivers’ quality of life and identify associated risk factors. Methods: A cross-sectional study was conducted from 21 January 2024 to 21 January 2025. A total of 117 primary caregivers of children with refractory epilepsy completed the World Health Organization Quality of Life (WHOQOL)-BREF (26 items) and Adherence questionnaire (6 items). Participants were divided into KD therapy groups (n = 51) and non-KD therapy groups (n = 66) according to the treatment. Factors associated with caregivers’ QoL in the ketogenic treatment were analyzed using the multifactor hierarchical regression. Results: There was no significant difference in QoL scores between the KD and non-KD caregiver groups (p > 0.05). KD adherence emerged as independently associated with caregivers’ QoL, particularly in the environmental domain (Model 1: β = −0.309, p = 0.022; Model 2: β = −0.306, p = 0.025). A higher KD cost was significantly associated with a lower social domain score in both models (Model 1: β = −0.285, p = 0.032; Model 2: β = −0.286, p = 0.034). Model 1 for the environmental domain demonstrated modest explanatory power (Adjusted R² = 0.246, p = 0.002). Conclusions: These findings underscore the need for clinical support systems to assess and address modifiable stressors early in treatment, including family structure, challenges with ketogenic diet therapy adherence, and financial burden. Such comprehensive evaluation is essential for developing effective and personalized interventions. Full article

The pervasive integration of artificial intelligence across e-commerce ecosystems has fundamentally transformed the competitive landscape, rendering systematic and reproducible platform evaluation frameworks an operational necessity rather than an academic exercise. Conventional multi-criteria decision analysis approaches for e-commerce evaluation remain structurally constrained by their dependency on human expert panels, which introduce recruitment costs, cognitive biases, limited reproducibility, and the practical infeasibility of assembling genuinely multidisciplinary panels spanning e-commerce strategy, machine learning engineering, and financial technology simultaneously. This study proposes a novel decision support framework that integrates Large Language Model (LLM) multi-agent simulation with the Best–Worst Method (BWM) to derive reproducible priority weights for AI-driven e-commerce platform evaluation within a rigorous business intelligence architecture. Twelve domain-differentiated LLM agents—organized into three expertise groups representing e-commerce management, AI and machine learning technology, and digital payment systems—were instantiated with structured system prompts encoding professional domain knowledge and deployed across three independent simulation rounds to perform BWM pairwise comparisons across a comprehensive six-dimensional, 30-sub-criterion evaluation hierarchy. Inter-agent consensus was synthesized through geometric mean aggregation, with consistency verification conducted via BWM’s xi* indicator and inter-round stability assessed through coefficient of variation analysis. Results reveal that Transaction Security and Trust achieves the highest dimension-level weight (w = 0.248), followed by AI Recommendation Effectiveness (w = 0.213), with Personal Data Protection (G = 0.0750), Recommendation Accuracy (G = 0.0607), and Transaction Transparency (G = 0.0549) emerging as the three highest globally ranked sub-criteria. The aggregated consistency indicator xi* = 0.062 confirms logical coherence of the multi-agent judgment consensus, and all dimension weights exhibit CV values below 2.8%, demonstrating exceptional inter-round stability. Spearman rank correlations among the three domain-expertise groups exceed 0.92, confirming strong inter-group convergence. Sensitivity analysis under perturbations of ±10% and ±20% demonstrates that the top-five priority indicators are structurally stable. This study establishes LLM multi-agent BWM simulation as a methodologically rigorous, institutionally accessible, and computationally reproducible alternative to traditional expert elicitation for complex platform evaluation tasks. Full article

Background: Quality of life (QoL) after spinal cord injury (SCI) is influenced by psychosocial factors, yet less is known about how these factors are examined across national contexts. Objective: This scoping review mapped studies examining depression, employment, and social participation in relation to QoL or health-related QoL (HRQoL) among individuals with SCI in the United States and South Korea. Methods: Following PRISMA-ScR guidelines, five databases were searched for peer-reviewed English- and Korean-language studies published between 2007 and 2025. Results: Sixteen studies were included: nine from South Korea and seven from the United States. Depression and psychological distress were associated with lower QoL/HRQoL in both countries, although South Korean studies more often examined depression with stress and functional concerns, whereas U.S. studies situated depression within participation, spirituality, and youth psychosocial functioning. Employment was linked to QoL/HRQoL in both contexts, with South Korean studies emphasizing economic activity, vocational rehabilitation, and financial strain, and U.S. studies emphasizing employment status and vocational outcomes. Social participation was important in both countries, but South Korean studies focused more on community transition, functional independence, and social attitudes, whereas U.S. studies emphasized participation contexts, accessibility, and social relationships. Conclusions: Across the three domains, depression, employment, and social participation emerged as recurring psychosocial domains associated with QoL/HRQoL after SCI in both countries. These differences suggest that psychosocial adaptation after SCI should be understood within cultural and rehabilitation contexts. Full article

We propose a transformer-based neural network for predicting credit card default using raw multivariate credit data represented as a 2D time series, eliminating the need for manual feature engineering. Unlike existing state-of-the-art (SOTA) tree-based models that rely heavily on handcrafted features, our model leverages self-attention to extract latent temporal patterns directly from the raw data. Evaluated on two real-world datasets, our approach outperforms the popular LightGBM baselines and achieves performance on par with the leading ensemble methods. To further explore if our proposed model can enhance common ensemble methods, we incorporate it into an ensemble together with LightGBM. Experimental results show that the ensemble integrating our proposed transformer-based model outperforms existing ensemble approaches. Designed with deployment in mind, the model architecture is lightweight, generalizable, and maintainable, making it suitable for integration into real-world credit risk pipelines. Our results demonstrate strong practical relevance and a clear path towards scalable deployment in financial applications. In addition, we have built in an optional feature augmentation extension to the proposed model to facilitate hybrid adoption of our model by existing users who are accustomed to engineered features from domain expertise and industry practice. Hence, our model is user-friendly and can leverage hybrid learning to support both user-crafted and model-learned features to improve model performance and deployment. Full article

This study aims to investigate the integration of generative artificial intelligence (GAI) and advanced large language models (LLMs) in financial sentiment research, focusing on improving the accuracy and robustness of financial sentiment classification from investor-generated textual data. The research employs advanced large language models, including XLNet, FinBERT, T5, Gemma-7B, Llama-2, and Llama-3, specifically fine-tuned to address the intricacies of financial language. We utilize generative AI models, such as GPT-4, GPT-3.5, and GPT-2, for data augmentation to mitigate scarcity. The fine-tuned Gemma-7b model proved to be the most successful, with a greater Success Rate (S-rate). The Gemma-7b model showed significant enhancements in performance after fine-tuning, highlighting its capacity to grasp the intricacies of financial emotion. This methodology provides a robust framework for financial sentiment classification and supports the extraction of meaningful sentiment signals from financial text. The results demonstrate the effectiveness of advanced LLMs for financial sentiment analysis and highlight their potential for supporting future research and analytical applications in financial text mining. Full article

Background/Objectives: Mental health problems among university students represent a growing public health concern and are shaped by both psychological and socioeconomic determinants that may act independently and interactively. This systematic review aimed to evaluate the separate and combined effects of these determinants on depression, anxiety, stress, and psychological distress in higher education students. Methods: A structured and targeted search strategy using predefined keyword groups and Boolean combinations across PubMed, Scopus, Web of Science, and Google Scholar identified 99 records, of which 19 duplicates were removed. After screening 80 titles and 52 abstracts, 34 full-text articles were assessed for eligibility, and 30 studies were ultimately included in the final review. Data were extracted on study characteristics, mental health outcomes, psychological determinants, socioeconomic factors, and their interactions. Results: The included studies consistently showed that psychological factors, including resilience, coping strategies, loneliness, self-efficacy, and perceived control, were associated with mental health outcomes, with higher resilience and self-efficacy linked to lower levels of depression and anxiety, and maladaptive coping and loneliness associated with increased psychological distress. Socioeconomic determinants, including financial stress, low socioeconomic status, parental education, housing insecurity, and food insecurity also independently contributed to elevated risks of depression, anxiety, and stress. Importantly, several studies demonstrated an interaction between these domains, where socioeconomic disadvantage amplified the adverse effects of poor coping capacity, low resilience, and social isolation, whereas social support and adaptive coping mitigated these effects. Conclusions: Student mental health is influenced by both distinct and interacting psychological and socioeconomic mechanisms, emphasizing the need for integrated institutional strategies that address structural vulnerabilities alongside individual psychological resilience. Full article

This review examines the evolution and architectural foundations of agentic artificial intelligence (AI), with a focus on collaborative multi-agent systems for complex task execution. The paper analyzes the core components, agent architectures, coordination mechanisms, application domains, and deployment challenges that enable autonomous reasoning and decision-making in real-world environments. To complement the survey, a comparative cryptocurrency market analysis case study is conducted using CrewAI, LangChain, and LangGraph focusing on workflow orchestration characteristics such as tool invocation, task transitions, orchestration depth, and memory integration. The findings are further supported by evidence from real-world financial applications reported in the literature, indicating productivity gains of 50–80% in financial data tasks and up to 20% improvement in stock prediction accuracy, highlighting the growing impact of multi-agent AI systems in market intelligence. The study highlights how architectural design choices influence reasoning continuity, coordination behavior, scalability, and system reliability, providing practical guidance for the design and deployment of agentic AI systems in complex, data-intensive domains. Full article

Digital technologies provide effective tools for formulating sustainable, evidence-based policies; however, this field has so far lacked a cohesive and practical framework to guide their application. Providing comprehensive answers to six primary research questions, this study aims to address this critical gap concerning the prerequisites, challenges, opportunities, key technologies, policy areas, and critical success factors (CSFs) for applying digital technologies in environmental sustainability policymaking. In this study, 39 articles were analyzed from 293 documents indexed in the Web of Science as of 19 August 2025, in accordance with the PRISMA 2020 guidelines. The prerequisites are categorized into the following themes: fiscal incentives, a culture of innovation and sustainability, effective regulations, robust digital infrastructures, participation, and reliable and accessible data. We identified significant challenges, including financial constraints, human resource deficits, infrastructural and regulatory gaps, and the adverse environmental impacts of digital technologies themselves. Opportunities emerged under two main domains: effective policymaking and enhanced environmental management. Our study indicates that pioneering technologies at the core of this transformation include artificial intelligence, big data, blockchain, the Internet of Things, machine learning, and robots. Their applications are predominant in key policy areas, including the environment, energy, climate change, urban sustainability, agriculture, industry, and food security. The analysis identifies four CSFs: the policy–digital–sustainability nexus, fundamental processes, soft capacities, and hard capacities. Full article

Existing Cyber Resilience Assessment Guidelines, including those of the Bank of Korea (BoK), focus on governance-oriented compliance and lack quantitative criteria for measuring the operational effectiveness of security technologies—a Policy–Technology Gap also common in general enterprise settings. To address this gap, this study proposes D3-CREF, a technology-centric cyber resilience evaluation framework that maps the MITRE D3FEND taxonomy to financial security domains and introduces a Normalized Resilience Index (NRI) aggregating four dimensions—Coverage, Maturity, Automation, and Timeliness—via a closed-form weighted geometric mean with AHP-elicited weights (consistency ratio CR = 0.04). All NRI indicators are anchored to MITRE ATT&CK techniques and exemplar CVE entries, enabling threat-informed measurement. The framework was validated through a three-round Delphi study with 50 experts (Kendall’s W = 0.78, p < 0.001; Cronbach’s α = 0.89; CVR 0.68–0.92) and a Cyber Range-based simulation. For three institutions with identical BoK scores (92/100), NRI yielded discriminative values of 0.83, 0.44, and 0.09 (CV = 0.68 vs. 0.00 for the baseline), confirming a shift from compliance-based to performance-driven assessment. Full article

Human-Centered Artificial Intelligence (HCAI) has emerged as a promising paradigm to increase transparency, usability, and trust in AI-driven Decision Support Systems (DSS). However, existing research lacks technically detailed accounts of how HCAI principles can be operationalized, implemented, and empirically validated in real decision environments. This study proposes a technically grounded HCAI-oriented DSS framework and presents a concrete prototype implemented in two high-stakes domains: clinical decision support and financial risk assessment. The architecture integrates interpretable machine learning models, SHAP-based explanations, structured user-feedback loops, and governance mechanisms aligned with the EU Trustworthy AI Guidelines. We trained and evaluated domain-specific models using publicly available medical and financial datasets, describing all data preprocessing, model selection, and hyperparameter settings to ensure reproducibility. An empirical study involving 30 domain experts (15 clinicians, 15 financial analysts) compared the HCAI-DSS with a functionally identical black-box DSS. Statistical analyses (paired t-tests with 95% confidence intervals and Cohen’s d) revealed that the HCAI-DSS significantly improved trust (d = 1.23), transparency and understanding (+1.76 mean difference), usability (SUS difference = +15.4), and decision accuracy (+10.2%), without a significant increase in decision time (p = 0.08). Qualitative feedback further demonstrated that explanations, control, and human-in-the-loop features increased confidence and reduced uncertainty. The results provide empirical evidence that HCAI principles tangibly enhance DSS effectiveness and user acceptance. The study contributes (1) a reproducible technical implementation, (2) a validated HCAI-DSS architecture, and (3) multi-domain evidence of improved decision quality. These findings support sustainable and trustworthy AI adoption across sectors and align with emerging regulatory frameworks such as the EU AI Act. Full article

Purpose: This study aims at the modification and psychometric evaluation of the “revised Illness Perception Questionnaire” (IPQ-R) for occupational dermatological rehabilitation. Methods: First, the questionnaire was modified for application in occupational dermatology. Subsequently, 254 patients of an inpatient rehabilitation programme participated in a cross-sectional survey. Afterwards, the dimensional analysis of the IPQ-R was conducted using principal component analysis. Separate analyses were conducted for the illness representations and the causal attribution scale. Results: A total of 228 participants were included in the analysis (age: M = 48.2 years; SD = 12.0; 53.9% female). The patient acceptance of the questionnaire was high (response rate 87.3%; rate of completion between 92.5% and 98.4%, N = 254). The IPQ-R for occupational dermatology consists of 29 items in the domain of illness representations, which include seven factors (illness coherence, emotional representations, consequences: implications for the structuring of own life, consequences: financial and social impacts, treatment control, personal control, and timeline acute/chronic). Six of these scales have acceptable-to-good internal consistency (Cronbach’s α 0.72–0.84); for one scale, the internal consistency is Cronbach’s α = 0.66. A separate analysis of the causes resulted in eight factors (psychological causes at work and during leisure time, attributions outside the workplace, skin cleansing and skin protection measures, behaviour-related risk factors, causes at work, other risk factors, external factors that cannot be influenced by the person, and climatic influences) with a total of 30 items. Five of the eight scales have an acceptable-to-good internal consistency (Cronbach’s α 0.71–0.83), and three scales are just below the acceptable range (Cronbach’s α 0.63–0.66). Conclusion: Overall, the initial psychometric results of the IPQ-R for occupational dermatology were satisfactory. However, additional validation steps are still required. The following differences to the original model should be considered when interpreting the available results: the factor “timeline cyclical” could not be replicated in this field of application. Additionally, two factors with different thematic emphases in the “consequences” section, besides effects on the personal way of life, social and financial consequences, became visible as well. Full article

The proliferation of digital financial news offers unprecedented opportunities for automated analysis of market-moving events. This paper presents a framework for event-based sentiment analysis of financial news that leverages Large Language Models (LLMs). The approach brings together three complementary ideas: Retrieval-Augmented Generation (RAG) for contextual enhancement, Graph Neural Networks (GNNs) for modeling relationships between events, and a multi-agent ensemble for orchestrated reasoning. The methodology targets well-known difficulties in financial text processing, including domain-specific terminology, implicit event detection, and temporal reasoning, and it combines transformer-based event extraction with sentiment classification enhanced by external knowledge retrieval. We evaluate six model configurations on an aggregated corpus of 14,851 financial news samples. On the event-detection task, every configuration reaches a weighted F1-score of 100%; we show that this is a ceiling effect produced by a binary event/no-event formulation over a highly imbalanced dataset rather than evidence of a difficult problem being solved, and we discuss what it implies for how such systems should be evaluated. On three-way sentiment classification, the strongest configuration—the multi-agent ensemble—reaches 87.4% accuracy, narrowly ahead of a RoBERTa (Robustly Optimized BERT Pretraining Approach) baseline at 87.2%; however, because the gaps reported between models are small and we did not run significance testing, we report them as indicative rather than definitive. The GNN component is described as part of the proposed design, but it has not yet been validated experimentally, and we state this limitation explicitly. The framework produces interpretable, structured outputs suited to downstream use in algorithmic trading, risk assessment, and investment decision support, and the paper contributes a reusable financial NLP pipeline together with a candid account of where the current evidence is, and is not, convincing. Full article

Federated learning offers a promising approach for cross-institutional financial risk control modeling but encounters two key challenges in practice: feature space heterogeneity and low sample overlap rate. Current federated transfer learning methods often rely heavily on sufficient overlapping samples or explicit feature alignment. However, these approaches frequently result in negative transfer when enforced alignment is applied in highly heterogeneous environments. To address this issue, we propose FedVI, a novel federated transfer learning framework that integrates tabular-to-image conversion and meta-learning mechanisms. Moving beyond conventional methods that rely on sample-level alignment, FedVI employs a federated dual-stream feature alignment strategy to securely reconstruct a unified global feature map across institutions. Subsequently, FedVI integrates federated Image Generator for Tabular Data (IGTD) with tabular Transformer technology to convert one-dimensional tabular data into two-dimensional visual-semantic tensors. These tensors effectively fuse spatial topology and semantic information while embedding an independent Mask channel to explicitly retain the true missingness patterns of features. Finally, FedVI adopts the Model-Agnostic Meta-Learning (MAML) architecture to facilitate global parameter optimization. We evaluated FedVI on the real-world Lending Club credit dataset and Home Credit Default Risk datasets under highly heterogeneous federated settings (i.e., heterogeneous feature spaces across three clients and scarce overlapping samples). The results reveal that FedVI achieves competitive performance against advanced baselines such as FedProx, FedRep, and FedKT, particularly in recall and F1-Score. These findings indicate that FedVI can effectively support cross-domain adaptation under heterogeneous federated learning settings. Full article