Frontiers in Artificial Intelligence最新文献

Purposes: The study investigated the intention of account executives from Small and Medium Enterprises (SMEs) to employ artificial intelligence at their workplace. This study will examine the Unified Theory of Acceptance and Use of Technology (UTAUT), as well as technological and personal characteristics, and the role of SME account executives in adopting artificial intelligence. This study addresses the knowledge gaps in SME account executives' understanding of artificial intelligence.

Methodology: employed an online questionnaire distributed in collaboration with SMEs in Malaysia to gather responses from 273 account executives who work in SMEs. The data were analyzed using PLS-SEM and Artificial Neural Networks (ANN) to investigate SME account executives' intentions to employ artificial intelligence. The demographic information of the individuals was analyzed using SPSS software.

Results: The study's findings revealed positive and significant relationships between performance expectancy, effort expectancy, social influence, facilitating conditions, system quality, employee awareness, and personal innovativeness toward artificial intelligence. Insignificant relationships were found between time-saving features and technological self-efficacy, and a negative, significant relationship existed with internet technology (IT) features toward artificial intelligence.

Limitation: The cross-sectional approach focuses on SMEs in Malaysia, where the study's applicability to other industries and countries is limited due to changes in the cultural, economic, and regulatory environment. Because participants may give socially acceptable answers rather than honest ones, using self-reported data raises the possibility of bias. Because inquiry assumes a certain level of knowledge with AI technology, respondents' varying levels of digital competency may influence the findings.

Practical implication: The findings of this study can help SMEs adopt artificial intelligence for their operations, particularly in accounting departments. Collaboration among organizations can help improve employee motivation to increase intention to use artificial intelligence.

Originality/value: This study uses the Unified Theory of Acceptance and Use of Technology (UTAUT), technical qualities, and individual traits.

Context: Fraud and corruption are among the main crimes affecting public institutions, with the healthcare sector being particularly vulnerable due to its structural complexity, the coexistence of public and private providers, the large number of actors involved, the globalized nature of supply chains, the high financial costs, and the information asymmetry among stakeholders. These factors weaken healthcare systems, resulting in resource waste, reduced resilience during medical emergencies, and limited access to essential services.

Objective: This study aims to evaluate automatic text summarization methods by comparing the quality of machine-generated summaries with those produced by humans, from the perspective of Data Scientists and SUS Auditors, within the context of audits carried out by the National Department of Unified Health System (Sistema Único de Saúde-SUS) Auditing (AudSUS).

Method: A controlled experiment was conducted to assess the performance of Small Language Models (SLMs) in summarization tasks, using the metrics ROUGE-N, ROUGE-L, BLEU, METEOR, and BERTScore. In addition, the consistency of results across 35 runs, their contribution to reducing information overload, and their pairwise performances were evaluated.

Results: The models NousResearch/Hermes-3-Llama-3.2-3B, Qwen/Qwen2.5-7B-Instruct, and meta-llama/Llama-3.2-3B-Instruct achieved the highest average performances across all metrics, standing out for their ability to preserve contextual meaning and synthesize essential information more effectively than human-generated summaries.

Conclusion: The findings highlight the potential of SLMs as tools to reduce information overload, thereby enhancing the effectiveness of the analytical phase of audits and enabling faster preparation of teams for the operational stage.

Natural Language Processing (NLP) has an influence on almost every field nowadays, such as business, healthcare, and sports, by making advanced interactions with human language and providing analytics. In the field of business, NLP has been a revolution, bettering customer service with the help of advanced chatbots, sentiment analysis, and automation in generating content, which enhances efficiency, personalization, and most importantly, decision-making. In healthcare, NLP is of crucial importance in decoding unstructured data like of medical records, supporting diagnostic accuracy, and making patient communication smoother, leading to better outcomes and improving efficiency. When it comes to sports, NLP provides critical insights through performance analytics, media content interpretation, and improved fan engagement, transforming data to utilize it for our advantage. The aim of this review is to systematically evaluate NLP's effectiveness across these sectors, address possible and existing challenges, and propose approaches for future research. Through the integration of case studies and performance assessments, we seek to clearly explain how NLP promotes innovation, resolves complex issues, and has made contributions to advancing new heights in these domains.

Tourism in biodiversity-rich regions was among the sectors most severely disrupted by the COVID-19 pandemic, which amplified existing socioeconomic vulnerabilities and placed cultural and natural heritage conservation at risk. In the Peruvian Amazon, Bagua Province illustrates this challenge, where experiential tourism is central to local livelihoods yet lacks adaptive management tools to support a sustainable recovery. To address this gap, this study introduces an integrated approach that combines artificial intelligence with biodiversity conservation through the application of multilayer perceptron (MLP) neural networks. By analyzing two decades of domestic visitor data (2003-2023), the research explores how predictive modeling can inform tourism governance in fragile ecosystems. Two scenarios were evaluated: one incorporating the complete dataset and another excluding the anomalous year 2020, heavily disrupted by the pandemic. The findings show that MLP models are capable of capturing visitor dynamics and forecasting demand fluctuations with notable accuracy. This predictive capacity allows for more adaptive planning of ecologically sensitive sites, such as the Tsunsuntsa Waterfall, where balancing visitor inflows with ecological thresholds is essential to preventing overtourism. Beyond technical accuracy, the study highlights the strategic potential of artificial intelligence as a governance tool that strengthens resilience in post-pandemic contexts, offering actionable insights for harmonizing socioeconomic recovery with biodiversity preservation. By positioning neural networks as vital instruments for sustainable destination management, this research contributes a reproducible model that can be adapted to other vulnerable regions worldwide. It underscores the value of integrating advanced computational methods into tourism governance frameworks, ultimately bridging technology and conservation to foster long-term sustainability.

To enable rapid, accurate grading of tunnel surrounding rock during construction, we propose a real-time grading method that integrates image processing with lightweight deep learning. We developed an automated pipeline that combines image-processing techniques and machine-learning algorithms to extract and classify characteristic parameters of tunnel surrounding rock, enabling real-time monitoring and classification at the tunnel palm surface. The study demonstrates that: (1) Following the proposed image-acquisition standards for rock and tunnel palm surfaces, images are converted to grayscale, denoised, enhanced, and normalized, which facilitates efficient and accurate extraction of structural features and improves the precision of classification parameters; (2) An optimized lithology identification and classification model was built, and a rock-hardness, strength, and integrity sensing approach based on the ShuffleNetV2 convolutional neural network was introduced to achieve real-time surrounding-rock grading. On an engineering site, the method attains 85% accuracy for lithology classification, 75% for rock-mass integrity, and 80% for overall surrounding-rock grade, confirming its feasibility and practical value. These results offer theoretical insight and engineering utility for the scientific evaluation of tunnel surrounding-rock grade.

Objective: Although imaging and paraspinal muscle parameters are linked to postoperative recurrent lumbar disc herniation (PRLDH), micro-level texture characteristics and their interactions remain underexplored. This study applied deep learning (DL)-radiomics to quantify the microstructural heterogeneity of responsible intervertebral discs and paraspinal muscles (L3-S1), and assessed a combined disc-muscle model for predicting PRLDH.

Method: Clinical and imaging data from 170 lumbar disc herniation (LDH) patients undergoing percutaneous transforaminal endoscopic surgery (Jan 2022-Dec 2024) were retrospectively analyzed. DL and radiomics features were extracted from intervertebral discs and paraspinal muscles. Feature selection via mutual information was followed by construction of a DL-radiomics Radscore model. Internal validation used leave-one-out, 10-fold cross-validation, and bootstrapping. Pfirrmann grading performance was compared with the disc Radscore, and potential disc-muscle interactions were explored using optimal cutoffs.

Results: Among 170 patients, 39 had postoperative recurrence. Disc Radscore included 2 DL and 3 radiomics features, while muscle Radscore comprised 2 DL and 5 radiomics features. The disc Radscore demonstrated good predictive ability (AUC 0.857, 95% CI 0.797-0.918) across validation methods (AUC 0.846-0.857). Muscle Radscore showed moderate performance (AUC 0.718, 95% CI 0.627-0.809). Pfirrmann grade poorly predicted recurrence (AUC 0.506, 95% CI 0.412-0.600). Combined disc-muscle analysis was less stable than disc Radscore alone.

Conclusion: DL-radiomics-derived intervertebral disc Radscore robustly predicts PRLDH. While combined disc-muscle assessment is less consistent, their interactions may inform postoperative risk stratification and management in LDH patients.

Introduction: Accurate segmentation of kidney masses and structure is essential for medical application including diagnosis and treatment. This research proposed the dual track hybrid VHUCS-Net architecture which effectively highlights structural size-shape variants, boundaries and complex structural features in kidney disease.

Methods: Efficient segmentation is achieved by integrating the transformer enhanced U-Net model with the contrast optimized Protuberance Detection Network (PDN) model. The process begins with analysing kidney images using a standard U-Net combined with Vision Transformer attention and a High Resolution Network (HRNet) which capture global dependencies while preserving high resolution features resulting in accurate segmentation of the kidney region. Also, the masked kidney image undergoes processing through a contrast optimized PDN model with multi scale pooling, contrast enhancement, boundary refinement and explicit feature fusion to segment the mass region thereby enhancing mass localization improving border identification and enabling accurate abnormality detection. The resulting features are fused to provide a refined mass segmentation result that exactly identifies the location and structural abnormalities.

Results: The VHUCS-Net model was evaluated using the kidney segmentation dataset achieving an intersection over union score of 0.9441 and a dice coefficient of 0.9712 showing outstanding segmentation precision.

Discussion: These results indicate improved diagnostic efficiency and support clinical decision making by providing more accurate and interpretable segmentation outputs. Moreover, VHUCS-Net is validated with additional publicly available datasets with image mask correspondence, therefore proving the model effectiveness and generalizability across many segmentation tasks. The results highlight the capability of the proposed VHUCS-Net model to enhance diagnostic accuracy and assist clinical decision making through more detailed and interpretable segmentation outcomes.

Artificial intelligence (AI) has shown promise in supporting clinical decision making, yet adoption in healthcare remains limited by concerns regarding transparency, verifiability, and accountability of AI-generated recommendations. In particular, generative and data-driven CDS systems often provide outputs without clearly exposing the evidentiary basis or reasoning process underlying their conclusions. This article presents a conceptual framework for auditable and source-verified AI-based clinical decision support, grounded in principles from evidence-based medicine, data provenance, and trustworthy AI. The proposed architecture integrates a curated medical knowledge base with explicit provenance metadata, a retrieval-augmented reasoning (RAG) engine that links generated recommendations to identifiable clinical guidelines and peer-reviewed sources, and a tamper-evident audit logging mechanism that records system inputs, retrieved evidence, and inference steps for retrospective review. This work does not introduce a new algorithm nor report a prototype implementation; rather, it synthesizes existing technical approaches into a coherent system design intended to improve traceability, clinician trust, and regulatory readiness. Key feasibility challenges are discussed, including knowledge-base governance and updating, citation fidelity in RAG architectures, bias propagation from underlying evidence, latency and usability trade-offs, privacy considerations, and alignment with emerging regulatory frameworks such as FDA Software as a Medical Device guidance and the European Union Artificial Intelligence Act. The article concludes by outlining a staged evaluation roadmap involving simulation studies and clinician-centered user research to guide future implementation and empirical validation.

Introduction: This article studies a new clustering-based federated learning algorithm that leverages Kullback-Leibler (KL) divergence to tackle heterogeneous data in wireless sensing environments.

Methods: Firstly, highdimensional heterogeneous data is subjected to principal component analysis to generate dimension-reduced representations, thereby reducing computational complexity. Secondly, the KL divergence distances between each pair of clients are calculated, followed by clustering according to the minimum threshold. The new KL divergence distance between the aggregated clients and others is taken as the average of the two. Finally, the federated learning training is conducted within each cluster to obtain a personalized model based on the classic wireless datasets.

Results and discussion: After the personalized models are tested, clients are reclustered and the models are updated-that is, a series of iterative operations, the optimal number of clusters and recognition accuracy are obtained. The test results show that the proposed algorithm based on KL divergence has higher recognition accuracy than several reported ones.

Implicit aspect detection aims to identify aspect categories that are not explicitly mentioned in text, but existing models struggle with four persistent challenges: aspect ambiguity, where multiple latent aspects are implied by the same expression, data imbalance and sparsity of implicit cues, contextual noise and syntactic variability in unstructured user reviews, and aspect drift, where the relevance of implicit cues changes across sentences or domains. To address these issues, this paper proposes the Transformer-Enhanced Graph Aspect Analyzer (TEGAA), a unified framework that tightly integrates dynamic expert routing, semantic representation refinement, and hierarchical graph reasoning. First, a Dynamic Expert Transformer (DET) equipped with a Dynamic Adaptive Expert Engine (DAEE) mitigates syntactic complexity and contextual noise by dynamically routing tokens to specialized expert sub-networks based on contextual and syntactic-semantic cues, enabling robust feature extraction for ambiguous implicit expressions. Second, Semantic Contrastive Learning (SCL) directly addresses data imbalance and weak implicit signals by enforcing semantic coherence among contextually related samples while increasing separability from irrelevant ones, thereby improving discriminability of sparse implicit aspect cues. Third, implicit aspect ambiguity and aspect drift are handled through a Graph-Enhanced Hierarchical Aspect Detector (GE-HAD), which models word- and sentence-level dependencies via context-aware graph attention. The incorporation of Attention Sinks prevents dominant but irrelevant tokens from overshadowing subtle implicit cues, while Pyramid Pooling aggregates multi-scale contextual information to stabilize aspect inference across varying linguistic scopes. Finally, an iterative feedback loop aligns graph-level reasoning with transformer-level expert routing, enabling adaptive refinement of aspect representations. Experiments on three benchmark datasets-Mobile Reviews, SemEval14, and Sentihood-demonstrate that TEGAA consistently outperforms state-of-the-art methods, achieving F1-scores above 0.88, precision above 0.89, recall above 0.87, accuracy exceeding 89%, and AUC values above 0.89. These results confirm TEGAA's effectiveness in resolving implicit aspect ambiguity, handling noisy and imbalanced data, and maintaining robust performance across domains.