Frontiers in Artificial Intelligence最新文献

{"title":"The implementation of artificial intelligence in upper extremity surgery: a systematic review.","authors":"Dylan Parry, Brennon Henderson, Paul Gaschen, Diane Ghanem, Evan Hernandez, Anceslo Idicula, Tammam Hanna, Brendan MacKay","doi":"10.3389/frai.2025.1621757","DOIUrl":"https://doi.org/10.3389/frai.2025.1621757","url":null,"abstract":"<p><strong>Introduction: </strong>The rapid expansion of artificial intelligence (AI) in medicine has led to its increasing integration into upper extremity (UE) orthopedics. The purpose of this systematic review is to investigate the current landscape and impact of AI in the field of UE surgery.</p><p><strong>Methods: </strong>Following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, a systematic search of PubMed was conducted to identify studies incorporating AI in UE surgery. Review articles, letters to the editor, and studies unrelated to AI applications in UE surgery were excluded.</p><p><strong>Results: </strong>After applying inclusion/exclusion criteria, 118 articles were included. The publication years ranged from 2009 to 2024, with a median and mode of 2022 and 2023, respectively. The studies were categorized into six main applications: automated image analysis (36%), surgical outcome prediction (20%), measurement tools (14%), prosthetic limb applications (14%), intraoperative aid (10%), and clinical decision support tools (6%).</p><p><strong>Discussion: </strong>AI is predominantly utilized in image analysis, including radiograph and MRI interpretation, often matching or surpassing clinician accuracy and efficiency. Additionally, AI-powered tools enhance the measurement of range of motion, critical shoulder angles, grip strength, and hand posture, aiding in patient assessment and treatment planning. Surgeons are increasingly leveraging AI for predictive analytics to estimate surgical outcomes, such as infection risk, postoperative function, and procedural costs. As AI continues to evolve, its role in UE surgery is expected to expand, improving decision-making, precision, and patient care.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1621757"},"PeriodicalIF":4.7,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12515950/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145293954","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of the accuracy and repeatability of Deepseek V3, Doubao, and Kimi1.5 in answering knowledge-related queries about chronic non-bacterial osteitis.","authors":"Zhenxing Zhu, Jun Xie, Longxin Zhou, Chaoran Yang, Feng Li","doi":"10.3389/frai.2025.1629149","DOIUrl":"https://doi.org/10.3389/frai.2025.1629149","url":null,"abstract":"<p><strong>Background: </strong>There are significant differences in the diagnosis and treatment of chronic non-bacterial osteitis (CNO), and there is an urgent need for health education efforts to enhance awareness of this condition. Deepseek V3, Doubao, and Kimi1.5 are highly popular language models in China that can provide knowledge related to diseases. This article aims to investigate the accuracy and reproducibility of the responses provided by these three artificial intelligence (AI) language models in answering questions about CNO.</p><p><strong>Methods: </strong>According to the latest expert consensus, 16 questions related to CNO were collected. The three AI language models were separately asked these questions at three different times. The answers were independently evaluated by two orthopedic experts.</p><p><strong>Results: </strong>Among the responses of the three AI models to 16 CNO-related questions across three rounds of testing, only Doubao received \"Completely incorrect\" ratings (accounting for 6.25%) in the third round of scoring by Reviewer 2. During the answering process, Doubao had the shortest response time and provided the most words in its answers. In the first and third rounds of scoring by the first expert, Kimi scored the highest (3.938 ± 0.342, 3.875 ± 0.873), while in the second round, Doubao scored the highest (3.875 ± 0.5). In the second round of scoring by the second expert, Doubao received the highest score (3.812 ± 0.403). In the first and third rounds, Kimi1.5 received the highest score (3.812 ± 0.602, 3.812 ± 0.704).</p><p><strong>Conclusion: </strong>Deepseek V3, Doubao, and Kimi1.5 are capable of answering most questions related to CNO with good accuracy and reproducibility, showing no significant differences.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1629149"},"PeriodicalIF":4.7,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12515971/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145293953","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Redefining digital health interfaces with large language models.","authors":"Fergus Imrie, Paulius Rauba, Mihaela van der Schaar","doi":"10.3389/frai.2025.1623339","DOIUrl":"10.3389/frai.2025.1623339","url":null,"abstract":"<p><p>Digital health tools have the potential to significantly improve the delivery of healthcare services. However, their adoption remains comparatively limited due, in part, to challenges surrounding usability and trust. Large Language Models (LLMs) have emerged as general-purpose models with the ability to process complex information and produce human-quality text, presenting a wealth of potential applications in healthcare. Directly applying LLMs in clinical settings is not straightforward, however, as LLMs are susceptible to providing inconsistent or nonsensical answers. We demonstrate how LLM-based systems, with LLMs acting as agents, can utilize external tools and provide a novel interface between clinicians and digital technologies. This enhances the utility and practical impact of digital healthcare tools and AI models while addressing current issues with using LLMs in clinical settings, such as hallucinations. We illustrate LLM-based interfaces with examples of cardiovascular disease and stroke risk prediction, quantitatively assessing their performance and highlighting the benefit compared to traditional interfaces for digital tools.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1623339"},"PeriodicalIF":4.7,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12511092/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281224","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fluorescent marker prediction for non-invasive optical imaging in bovine satellite cells using deep learning.","authors":"Sania Sinha, Aarham Wasit, Won Seob Kim, Jongkyoo Kim, Jiyoon Yi","doi":"10.3389/frai.2025.1577027","DOIUrl":"10.3389/frai.2025.1577027","url":null,"abstract":"<p><p>Assessing the quality of bovine satellite cells (BSCs) is vital for advancing tissue engineered muscle constructs with applications in sustainable protein research. In this study, we present a non-invasive deep learning approach for optical imaging that predicts fluorescent markers directly from brightfield microscopy images of BSC cultures. Using a convolutional neural network based on the U-Net architecture, our method simultaneously predicts two key fluorescent signals, specifically DAPI and Pax7, which serve as biomarkers for cell abundance and differentiation status. An image preprocessing pipeline featuring fluorescent signal denoising was implemented to enhance prediction performance and consistency. A dataset comprising 48 biological replicates was evaluated using statistical metrics such as the Pearson <i>r</i> (correlation coefficient), the mean squared error (MSE), and the structural similarity Index (SSIM). For DAPI, denoising improved the Pearson <i>r</i> from 0.065 to 0.212 and SSIM from 0.047 to 0.761 (with MSE increasing from 9.507 to 41.571). For Pax7, the Pearson <i>r</i> increased from 0.020 to 0.124 and MSE decreased from 44.753 to 18.793, while SSIM remained low, reflecting inherent biological heterogeneity. Furthermore, enhanced visualization techniques, including color mapping and image overlay, improved the interpretability of the predicted outputs. These findings underscore the importance of optimized data preprocessing and demonstrate the potential of AI to advance non-invasive optical imaging for cellular quality assessment in tissue biology. This work also contributes to the broader integration of machine learning and computer vision methods in biological and agricultural applications.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1577027"},"PeriodicalIF":4.7,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12511074/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281239","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SMCFO: a novel cuttlefish optimization algorithm enhanced by simplex method for data clustering.","authors":"Kalpanarani K, Hannah Grace G","doi":"10.3389/frai.2025.1677059","DOIUrl":"10.3389/frai.2025.1677059","url":null,"abstract":"<p><strong>Introduction: </strong>In unsupervised learning, data clustering is essential. However, many current algorithms have issues like early convergence, inadequate local search capabilities, and trouble processing complicated or unbalanced input. Established methods like Kmeans are still widely used because of their ease of use; however, they struggle with non-spherical cluster shapes, which are sensitive to initialization, and suffer in highdimensional space. As a substitute, metaheuristic algorithms have surfaced as possible options, providing powerful global search ability. The Cuttlefish Optimization Algorithm (CFO) shows promise in clustering applications but suffers from premature convergence and poor local optimization capability.</p><p><strong>Methods: </strong>This paper introduces a new clustering method based on the Cuttlefish Optimization Algorithm (CFO), which improves upon the Nelder-Mead simplex method known as SMCFO. The method partitions the population into four subgroups with specific update strategies. One subgroup uses the Nelder-Mead method to improve the quality of solutions, while the others attempt to maintain exploration and exploitation equilibrium. This study compares the performance of the suggested SMCFO algorithm with four established clustering algorithms: CFO, PSO, SSO, and SMSHO. The evaluation used 14 datasets, which include two artificial datasets and 12 benchmark datasets sourced from the UCI Machine Learning Repository.</p><p><strong>Results and discussion: </strong>The proposed SMCFO algorithm consistently outperformed competing methods across all datasets, achieving higher clustering accuracy, faster convergence, and improved stability. The robustness of these outcomes was further confirmed through nonparametric statistical tests, which demonstrated that the performance improvements of SMCFO were statistically significant and not due to chance. The results confirm that the simplex-enhanced design boosts local exploitation and stabilizes convergence, which underlies SMCFO's superior performance compared to baseline methods.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1677059"},"PeriodicalIF":4.7,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12511114/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281256","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-scale and deeply supervised network for image splicing localization.","authors":"Sheng Qin, Ce Liang, Yuling Luo, Junxiu Liu, Qiang Fu, Xue Ouyang","doi":"10.3389/frai.2025.1655073","DOIUrl":"10.3389/frai.2025.1655073","url":null,"abstract":"<p><p>When maliciously tampered images are disseminated in the media, they can potentially cause adverse effects and even jeopardize national security. Therefore, it is necessary to investigate effective methods to detect tampered images. As a challenging task, the localization of image splicing tampering investigates whether an image contains tampered regions spliced from another image. Given the lack of global information interactions in existing methods, a multi-scale, deeply supervised image splicing tampering localization network is proposed. The proposed network is based on an encoder-decoder architecture, where the decoder uses different levels of feature maps to supervise the locations of splicing, enabling pixel-wise prediction of tampered regions. Moreover, a multi-scale feature extraction module is utilized between the encoder and decoder, which expands the global view of the network, thereby enabling more effective differentiation between tampered and non-tampered regions. F1 scores of 0.891 and 0.864 were achieved using the CASIA and COLUMB datasets, respectively; and the proposed model was able to accurately locate tampered regions.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1655073"},"PeriodicalIF":4.7,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507846/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281304","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integration of AI and ML in regenerative braking for electric vehicles: a review.","authors":"Zacharia Prakash","doi":"10.3389/frai.2025.1626804","DOIUrl":"10.3389/frai.2025.1626804","url":null,"abstract":"<p><p>Electric vehicle technology has grown rapidly in recent years due to battery advancements, environmental concerns and supportive policies. Regenerative braking systems play a critical role in improving energy efficiency by converting kinetic energy into electrical energy, thereby extending battery life and vehicle range. However, conventional regenerative braking faces challenges in energy recovery, comfort, and adaptability. Optimizing energy recovery ensures prolonged battery life by preventing overcharging or undercharging, making EVs more sustainable and cost-effective. This review paper explores the integration of Artificial Intelligence and machine learning techniques in regenerative braking systems to overcome these challenges. This study examines AI techniques such as regression models, neural networks, deep reinforcement learning, fuzzy logic, genetic algorithm and swarm intelligence based techniques for regenerative braking. The study also compares AI-based strategies with traditional braking methods. Unlike previous studies, which focus on individual AI techniques, this paper provides a comparative analysis of multiple AI approaches, assessing their impact on braking performance and energy recovery, and propose a hybrid AI framework. This paper covers challenges in real-time implementation, road adaptability, and vehicle control integration. This paper also discusses future research that optimize braking performance like V2X communication, edge computing, and explainable AI etc.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1626804"},"PeriodicalIF":4.7,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507741/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281247","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diversity-enhanced reconstruction as plug-in defenders against adversarial perturbations.","authors":"Zeshan Pang, Xuehu Yan, Shasha Guo, Yuliang Lu","doi":"10.3389/frai.2025.1665106","DOIUrl":"10.3389/frai.2025.1665106","url":null,"abstract":"<p><p>Deep learning models are susceptible to adversarial examples. In large-scale deployed services, plug-in defenders efficiently defend against such attacks. Plug-in defenders take two approaches to mitigate adversarial effects: input reconstruction and random transformations. Existing plug-in defense lacks diversity in transformation formulation due to the inherent feature preservation nature, which leads to vulnerability under adaptive attacks. To address this issue, we propose a novel plug-in defense named Diversity-enhanced Reconstruction (DeR). DeR counters adversarial attacks by frequency-aware reconstructors with enhanced diversity. Specifically, we design the reconstructors as a U-Net backbone with additional frequency components. The reconstructors are trained on the proposed DeR loss, which optimizes the reconstruction and diversity objectives jointly. Once trained, DeR can produce heterogeneous gradients and be applied as a plug-in defense. We conduct extensive experiments on three datasets and four classifier architectures under strict adversarial settings. The results demonstrate the superior robustness of DeR compared to state-of-the-art plug-in defense and the efficiency of DeR in real-time processing.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1665106"},"PeriodicalIF":4.7,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507835/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281251","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence as a tool to enhance social interventions in reducing crime.","authors":"Haya H Tarawneh, Niveen Halalsheh, Bahjat Abu Sulaiman, Huda Alhajjaj, Nesreen N Atieh","doi":"10.3389/frai.2025.1661266","DOIUrl":"10.3389/frai.2025.1661266","url":null,"abstract":"<p><strong>Introduction: </strong>This study explores the significant role of artificial intelligence (AI) in crime reduction, identifies the main challenges hindering its implementation, and examines differences in coping strategies between individuals in Jordan and Saudi Arabia.</p><p><strong>Methods: </strong>The research surveyed 170 AI professionals, equally divided between the two countries, with an average age of 45.2 years. Data were collected using a specially designed questionnaire assessing perceptions of AI, barriers to adoption, and coping mechanisms.</p><p><strong>Results: </strong>The findings indicated that AI plays a significant role in crime reduction. High levels of challenges were reported in implementing AI, and coping strategies related to AI in crime reduction were also assessed at a high level. No statistically significant differences were found in the level of challenges facing AI between Jordanian and Saudi participants.</p><p><strong>Discussion: </strong>This research contributes to understanding AI's practical applications in crime prevention and provides valuable insights for policymakers to strengthen AI adoption and overcome existing barriers in the region.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1661266"},"PeriodicalIF":4.7,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507826/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281293","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI competence and sentiment: a mixed-methods study of attitudes and open-ended reflections.","authors":"Gatis Lāma, Agnese Lastovska","doi":"10.3389/frai.2025.1658791","DOIUrl":"10.3389/frai.2025.1658791","url":null,"abstract":"<p><p>As artificial intelligence (AI) technologies become increasingly integrated into everyday life, understanding how the public perceives and interacts with AI is essential for fostering responsible and secure adoption. This study investigates the relationship between self-assessed AI competence, trust in AI-generated content, and sentiment toward AI among public and private sector employees in Latvia. Using a mixed-methods approach, the research combines quantitative survey data with open-ended qualitative responses to explore how demographic factors influence AI-related perceptions. Results reveal that although participants rate their AI competence and trust relatively highly, a significant portion of respondents either do not use AI or use it only for simple tasks. Sentiment toward AI is generally positive but often neutral, indicating that public attitudes are still forming. Statistically significant differences in AI competence were found across gender, age, and work sector, while trust in AI varied by education and age. Sentiment remained consistent across groups. Importantly, AI competence was positively correlated with trust, which in turn correlated with sentiment. Thematic analysis identified concerns about risk assessment, ethical implications, and the uncertain role of AI in daily life. The study underscores the need to enhance AI literacy and critical evaluation skills to ensure informed trust and societal resilience. These findings inform future strategies for public education, workforce training, and digital security policy in the context of accelerating AI adoption.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1658791"},"PeriodicalIF":4.7,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12504219/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145259457","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}