Sci Rep. 2025 May 23;15(1):18006. doi: 10.1038/s41598-025-96203-3.

ABSTRACT

In the contemporary technological world, fortifying cybersecurity defense against dynamic threat landscapes is imperative. Malware detectors play a critical role in this endeavor, utilizing various techniques such as statistical analysis, static and dynamic analysis, and machine learning (ML) to compare signatures and identify threats. Deep learning (DL) aids in accurately classifying complex malware features. The cross-domain research in data fusion strives to integrate information from multiple sources to augment reliability and minimize errors in detecting sophisticated cyber threats. This collaborative approach is the least addressed and pivotal for protecting against the advancing environment of modern malware attacks. This study presents a state-of-the-art malware analysis framework that employs a multimodal approach by integrating malware images and numeric features for effective malware classification. The experiments are performed sequentially, encompassing data preprocessing, feature selection using Neighbourhood Component Analysis (NCA), and dataset balancing with Synthetic Minority Over-sampling Technique (SMOTE). Subsequently, the late fusion technique is utilized for multimodal classification by employing Random Under Sampling and Boosting (RUSBoost) and the proposed ensemble deep neural network. The RUSBoost technique involves random undersampling and adaptive boosting to moderate bias toward majority classes while improving minority class (malware) detection. Multimodal Late fusion experimental results (95.36%) of RUSBoost (numeric) and the proposed model (imagery) outperform the standalone prevailing results for imagery (95.02%) and numeric (93.36%) data. The effectiveness of the proposed model is verified through the evaluation metrics such as Recall (86.5%), F1-score (85.0%), and Precision (79.9%). The multimodal late fusion of numeric and visual data makes the model more robust in detecting diverse malware variants. The experimental outcomes demonstrate that multimodal analysis may efficiently increase the identification strength and accuracy, especially when majority vote and bagging are employed for late fusion.

PMID:40410526 | DOI:10.1038/s41598-025-96203-3

Interdiscip Sci. 2025 May 23. doi: 10.1007/s12539-025-00714-6. Online ahead of print.

ABSTRACT

BACKGROUNDS: During the development of new drugs, it is essential to assess their effectiveness and examine the potential mechanisms behind side effects. This process typically involves combining the analysis of drugs under development with relevant existing drugs to more precisely evaluate the effects of drugs and targets. The use of deep learning methods to analyze this problem is currently a research hotspot, but several limitations remain: (i) how to deepen the analysis from the molecular level to the atomic level and analyze the key substructures that affect interactions on the basis of pharmaceutical mechanisms; (ii) how to integrate biomedical analysis with deep learning methods to make it medically sound and enhance interpretability.

METHODS: To address the limitations of existing research, based on Deep Graph Convolutional Network (Deep-GCN) and Bilinear Attention Network (BAN), we have constructed an interpretable deep learning framework, WDGBANDTI, to analyze and predict drug‒target interactions at the substructure level and enhance the prediction capability of the model with respect to unidentified target pairings by adding modules.

RESULTS: For different application scenarios, we validated the model via several commonly used and highly covered datasets. We also selected several state-of-the-art computer methods as comparison objects, and our model demonstrates advantages in accuracy, sensitivity, specificity, and other deep learning features. More importantly, the model can identify the substructures that play a role in drug‒target interactions through BAN, highlighting its excellent interpretability.

CONCLUSION: In conclusion, we believe that our work will contribute to advancements in drug development and side effect experiments and provide meaningful guidance for drug design.

PMID:40410523 | DOI:10.1007/s12539-025-00714-6

Nat Biotechnol. 2025 May 23. doi: 10.1038/s41587-025-02654-4. Online ahead of print.

ABSTRACT

The dominant success of deep learning techniques on protein structure prediction has challenged the necessity and usefulness of traditional force field-based folding simulations. We proposed a hybrid approach, deep-learning-based iterative threading assembly refinement (D-I-TASSER), which constructs atomic-level protein structural models by integrating multisource deep learning potentials with iterative threading fragment assembly simulations. D-I-TASSER introduces a domain splitting and assembly protocol for the automated modeling of large multidomain protein structures. Benchmark tests and the most recent critical assessment of protein structure prediction, 15 experiments demonstrate that D-I-TASSER outperforms AlphaFold2 and AlphaFold3 on both single-domain and multidomain proteins. Large-scale folding experiments further show that D-I-TASSER could fold 81% of protein domains and 73% of full-chain sequences in the human proteome with results highly complementary to recently released models by AlphaFold2. These results highlight a new avenue to integrate deep learning with classical physics-based folding simulations for high-accuracy protein structure and function predictions that are usable in genome-wide applications.

PMID:40410405 | DOI:10.1038/s41587-025-02654-4

Sci Rep. 2025 May 23;15(1):17921. doi: 10.1038/s41598-025-02491-0.

ABSTRACT

Remote sensing images (RSI), such as aerial or satellite images, produce a large-scale view of the Earth's surface, which gets them used to track and monitor vehicles from several settings, like border control, disaster response, and urban traffic surveillance. Vehicle detection and classification using RSIs is a vital application of computer vision and image processing. It contains locating and identifying vehicles from the image. It is done using many approaches that have object detection approaches, namely YOLO, Faster R-CNN, or SSD, which utilize deep learning (DL) to locate and identify the image. Additionally, the classification of vehicles from RSIs contains classification of them based on their variety, such as trucks, motorcycles, cars or buses, utilizing machine learning (ML) techniques. This article designed and developed an automated vehicle type detection and classification using a chaotic equilibrium optimization algorithm with deep learning (VDTC-CEOADL) on high-resolution RSIs. The VDTC-CEOADL technique presented examines high-quality RSIs for the accurate detection and classification of vehicles. The VDTC-CEOADL technique employs a YOLO-HR object detector with a residual network as the backbone model to accomplish this. In addition, CEOA based hyperparameter optimizer is designed for the parameter tuning of the ResNet model. For the vehicle classification process, the VDTC-CEOADL technique exploits the attention-based long-short-term memory (ALSTM) mod-el. Performance validation of the VDTC-CEOADL technique is validated on a high-resolution RSI dataset, and the results portrayed the supremacy of the VDTC-CEOADL technique in terms of different measures.

PMID:40410394 | DOI:10.1038/s41598-025-02491-0

Phys Rev Lett. 2025 May 2;134(17):176503. doi: 10.1103/PhysRevLett.134.176503.

ABSTRACT

Strong correlation brings a rich array of emergent phenomena, as well as a daunting challenge to theoretical physics study. In condensed matter physics, the fractional quantum Hall effect is a prominent example of strong correlation, with Landau level mixing being one of the most challenging aspects to address using traditional computational methods. Deep learning real-space neural network wave function methods have emerged as promising architectures to describe electron correlations in molecules and materials, but their power has not been fully tested for exotic quantum states. In this work, we employ real-space neural network wave function techniques to investigate fractional quantum Hall systems. On both 1/3 and 2/5 filling systems, we achieve energies consistently lower than exact diagonalization results which only consider the lowest Landau level. We also demonstrate that the real-space neural network wave function can naturally capture the extent of Landau level mixing up to a very high level, overcoming the limitations of traditional methods. Our work underscores the potential of neural networks for future studies of strongly correlated systems and opens new avenues for exploring the rich physics of the fractional quantum Hall effect.

PMID:40408749 | DOI:10.1103/PhysRevLett.134.176503

Phys Rev Lett. 2025 May 2;134(17):177301. doi: 10.1103/PhysRevLett.134.177301.

ABSTRACT

Transfer learning (TL) is a well-established machine learning technique to boost the generalization performance on a specific (target) task using information gained from a related (source) task, and it crucially depends on the ability of a network to learn useful features. Leveraging recent analytical progress in the proportional regime of deep learning theory (i.e., the limit where the size of the training set P and the size of the hidden layers N are taken to infinity keeping their ratio α=P/N finite), in this Letter we develop a novel single-instance Franz-Parisi formalism that yields an effective theory for TL in fully connected neural networks. Unlike the (lazy-training) infinite-width limit, where TL is ineffective, we demonstrate that in the proportional limit TL occurs due to a renormalized source-target kernel that quantifies their relatedness and determines whether TL is beneficial for generalization.

PMID:40408730 | DOI:10.1103/PhysRevLett.134.177301

PLoS One. 2025 May 23;20(5):e0321008. doi: 10.1371/journal.pone.0321008. eCollection 2025.

ABSTRACT

Accurate monthly runoff forecasting is vital for water management, flood control, hydropower, and irrigation. In glacierized catchments affected by climate change, runoff is influenced by complex hydrological processes, making precise forecasting even more challenging. To address this, the study focuses on the Lotschental catchment in Switzerland, conducting a comprehensive comparison between deep learning and ensemble-based models. Given the significant autocorrelation in runoff time series data, which may hinder the evaluation of prediction models, a novel statistical method is employed to assess the effectiveness of forecasting models in detecting turning points in the runoff data. The performance of Extreme Gradient Boosting (XGBoost) was compared with long short-term memory (LSTM) and random forest (RF) models for one-month-ahead runoff forecasting. The study used 20 years of runoff data (2002-2021), with 70% (2002-2015) dedicated for training and calibration, and the remaining data (2016-2021) for testing. The findings for the testing phase results show that the XGBoost model achieves the best accuracy, with R² of 0.904, RMSE of 1.554 m³/sec, an NSE of 0.797, and Willmott index (d) of 0.972, outperforming both the LSTM and RF models. The study also found that the XGBoost model estimated turning points more accurately, obtaining forecasting improvements of up to 22% to 34% compared to LSTM and RF models. Overall, the study's findings are essential for global water resource management, providing insights that can inform sustainable practices to support societies impacted by climate change.

PMID:40408639 | DOI:10.1371/journal.pone.0321008

PLoS Comput Biol. 2025 May 23;21(5):e1013071. doi: 10.1371/journal.pcbi.1013071. eCollection 2025 May.

ABSTRACT

Deep learning-based methods for identifying and tracking cells within microscopy images have revolutionized the speed and throughput of data analysis. These methods for analyzing biological and medical data have capitalized on advances from the broader computer vision field. However, cell tracking can present unique challenges, with frequent cell division events and the need to track many objects with similar visual appearances complicating analysis. Existing architectures developed for cell tracking based on convolutional neural networks (CNNs) have tended to fall short in managing the spatial and global contextual dependencies that are crucial for tracking cells. To overcome these limitations, we introduce Cell-TRACTR (Transformer with Attention for Cell Tracking and Recognition), a novel deep learning model that uses a transformer-based architecture. Cell-TRACTR operates in an end-to-end manner, simultaneously segmenting and tracking cells without the need for post-processing. Alongside this model, we introduce the Cell-HOTA metric, an extension of the Higher Order Tracking Accuracy (HOTA) metric that we adapted to assess cell division. Cell-HOTA differs from standard cell tracking metrics by offering a balanced and easily interpretable assessment of detection, association, and division accuracy. We test our Cell-TRACTR model on datasets of bacteria growing within a defined microfluidic geometry and mammalian cells growing freely in two dimensions. Our results demonstrate that Cell-TRACTR exhibits strong performance in tracking and division accuracy compared to state-of-the-art algorithms, while also meeting traditional benchmarks in detection accuracy. This work establishes a new framework for employing transformer-based models in cell segmentation and tracking.

PMID:40408631 | DOI:10.1371/journal.pcbi.1013071

PLoS One. 2025 May 23;20(5):e0324347. doi: 10.1371/journal.pone.0324347. eCollection 2025.

ABSTRACT

The agricultural sector faces critical challenges, including significant crop losses due to undetected plant diseases, inefficient monitoring systems, and delays in disease management, all of which threaten food security worldwide. Traditional approaches to disease detection are often labor-intensive, time-consuming, and prone to errors, making early intervention difficult. This paper proposes an intelligent framework for automated crop health monitoring and early disease detection to overcome these limitations. The system leverages deep learning, cloud computing, embedded devices, and the Internet of Things (IoT) to provide real-time insights into plant health over large agricultural areas. The primary goal is to enhance early detection accuracy and recommend effective disease management strategies, including crop rotation and targeted treatment. Additionally, environmental parameters such as temperature, humidity, and water levels are continuously monitored to aid in informed decision-making. The proposed framework incorporates Convolutional Neural Network (CNN), MobileNet-1, MobileNet-2, Residual Network (ResNet-50), and ResNet-50 with InceptionV3 to ensure precise disease identification and improved agricultural productivity.

PMID:40408612 | DOI:10.1371/journal.pone.0324347

Sci Adv. 2025 May 23;11(21):eadu2151. doi: 10.1126/sciadv.adu2151. Epub 2025 May 23.

ABSTRACT

Non-small cell lung cancer (NSCLC) constitutes over 80% of lung cancer cases and remains a leading cause of cancer-related mortality worldwide. Despite the advent of immune checkpoint inhibitors, their efficacy is limited to 27 to 45% of patients. Identifying likely treatment responders is essential for optimizing healthcare and improving quality of life. We generated multiplex immunofluorescence (mIF) images, histopathology, and RNA sequencing data from human NSCLC tissues. Through the analysis of mIF images, we characterized the spatial organization of 1.5 million cells based on the expression levels for 33 biomarkers. To enable large-scale characterization of tumor microenvironments, we developed NucSegAI, a deep learning model for automated nuclear segmentation and cellular classification in histology images. With this model, we analyzed the morphological, textural, and topological phenotypes of 45.6 million cells across 119 whole-slide images. Through unsupervised phenotype discovery, we identified specific lymphocyte phenotypes predictive of immunotherapy response. Our findings can improve patient stratification and guide selection of effective therapeutic regimens.

PMID:40408481 | DOI:10.1126/sciadv.adu2151

Genome Biol Evol. 2025 May 23:evaf101. doi: 10.1093/gbe/evaf101. Online ahead of print.

ABSTRACT

Plant DNA methylation changes occur hundreds to thousands of times faster than DNA mutations and can be transmitted transgenerationally, making them useful for studying population-scale patterns in clonal or selfing species. However, a state-of-the-art approach to use them for inferring population genetic processes and demographic histories is lacking. To address this, we compare evolutionary signatures extracted from CG methylomes and genomes in Arabidopsis thaliana and Brachypodium distachyon. While methylation variants (SMPs) are less effective than single nucleotide polymorphisms (SNPs) for identifying population differentiation in A. thaliana, they can classify phenotypically divergent B. distachyon subgroups that are otherwise genetically indistinguishable. The site frequency spectra generated using methylation sites from varied genomic locations and evolutionary conservation exhibit an excess of rare alleles. Nucleotide diversity estimates were three orders of magnitude higher for methylation variants than for SNPs in both species, driven by the higher epimutation rate. Correlations between SNPs and SMPs in nucleotide diversity and allele frequencies at gene exons are weak or absent in A. thaliana, possibly because the two sources of variation reflect evolutionary forces acting at different timescales. Linkage disequilibrium quickly decays within 100 bp for methylation variants in both plant species. Finally, we have developed a novel deep learning-based approach that infers demographic histories using methylation variation data alone. We identified recent population expansions in A. thaliana and B. distachyon using methylation variants that were not identified when using SNPs. Our study demonstrates the unique evolutionary insights methylomes provide that SNPs alone cannot reveal.

PMID:40408446 | DOI:10.1093/gbe/evaf101

PLoS One. 2025 May 23;20(5):e0318372. doi: 10.1371/journal.pone.0318372. eCollection 2025.

ABSTRACT

In the realm of game playing, deep reinforcement learning predominantly relies on visual input to map states to actions. The visual data extracted from the game environment serves as the primary foundation for state representation in reinforcement learning agents. However, humans leverage additional sensory inputs, such as audio cues, which play a pivotal role in perception and decision-making. Therefore, incorporating raw audio along with visual information shows potential for offering valuable insights to reinforcement learning agents. This study advocates for the integration of raw audio samples as complementary information to visual data in state representation. By using raw audio with visual cues, our objective is to enrich the decision-making process of the agent at each stage. Experimental evaluation were conducted employing Deep Q Networks (DQN) and Proximal Policy Optimization (PPO) algorithms within ViZDoom and Unity reinforcement learning environments. The results of our experiments reveal that augmenting visual information with raw audio samples yields superior rewards and expedites the learning rate compared to relying solely on visual data. Additionally, the findings suggest that considering both visual and audio features enhances the agent's behavior, a trend observed across Unity and ViZDoom environments. This study underscores the potential advantages of incorporating multisensory information, particularly raw audio, into the state representation of reinforcement learning agents. Such insights contribute to advancing our understanding of how agents perceive and engage with their environments, ultimately enhancing performance in complex gaming scenarios.

PMID:40408327 | DOI:10.1371/journal.pone.0318372

PLoS One. 2025 May 23;20(5):e0321179. doi: 10.1371/journal.pone.0321179. eCollection 2025.

ABSTRACT

The rapid proliferation of the Power Internet of Things (PIoT) has given rise to severe network security threats, with eavesdropping attacks emerging as a paramount concern. Traditional eavesdropping detection methods struggle to adapt to complex and dynamic attack patterns, necessitating the exploration of more intelligent and efficient anomaly localization approaches. This paper proposes an innovative method for eavesdropping node localization based on Kolmogorov-Arnold Networks (KANs). Leveraging the powerful ability of KANs to approximate arbitrary nonlinear functions, this method constructs an end-to-end mapping from heterogeneous node features to eavesdropping locations through flexible combinations of spline functions. To address the challenges of real-world power grid environments, this paper designs optimization strategies such as adaptive grid refinement and hierarchical sparsity regularization, further enhancing the model's robustness and interpretability. Extensive simulations and experiments on real power grid data demonstrate that the proposed method significantly outperforms traditional machine learning and mainstream deep learning approaches in terms of localization accuracy, generalization ability, and computational efficiency. This paper provides new perspectives and tools for intelligent power grid information security in IoT environments, holding significant innovative value in both theory and practice.

PMID:40408323 | DOI:10.1371/journal.pone.0321179

PLoS One. 2025 May 23;20(5):e0321856. doi: 10.1371/journal.pone.0321856. eCollection 2025.

ABSTRACT

The growing reliance on video conferencing software brings significant benefits but also introduces challenges, particularly in managing audio quality. In multi-participant settings, ambient noise and interruptions can hinder speaker recognition and disrupt the flow of conversation. This work proposes an audio-visual source separation pipeline designed specifically for video conferencing and telepresence robots applications. The framework aims to isolate and enhance the speech of individual participants in noisy environments while enabling control over the volume of specific individuals captured in the video frame. The proposed pipeline comprises key components: a deep learning-based feature extractor for audio and video, an audio-guided visual attention mechanism, a module for background noise suppression and human voice separation, and Deep Multi-Resolution Network (DMRN) modules. For human voice separation, the DPRNN-TasNet, a robust deep neural network framework, is employed. Experimental results demonstrate that the methodology effectively isolates and amplifies individual participants' speech, achieving a test accuracy of 71.88 % on both the AVE and Music 21 datasets.

PMID:40408322 | DOI:10.1371/journal.pone.0321856

Int J Comput Assist Radiol Surg. 2025 May 23. doi: 10.1007/s11548-025-03412-2. Online ahead of print.

ABSTRACT

PURPOSE: Stroke remains a leading cause of morbidity and mortality worldwide, despite advances in treatment modalities. Endovascular thrombectomy (EVT), a revolutionary intervention for ischemic stroke, is limited by its reliance on 2D fluoroscopic imaging, which lacks depth and comprehensive vascular detail. We propose a novel AI-driven pipeline for 3D CTA to 2D DSA cross-modality registration, termed DeepIterReg.

METHODS: The proposed pipeline integrates neural network-based initialization with iterative optimization to align pre-intervention and peri-intervention data. Our approach addresses the challenges of cross-modality alignment, particularly in scenarios involving limited shared vascular structures, by leveraging synthetic data, vein-centric anchoring, and differentiable rendering techniques.

RESULTS: We assess the efficacy of DeepIterReg through quantitative analysis of capture ranges and registration accuracy. Results show that our method can accurately register 70% of a test set of 20 patients and can improve capture ranges when performing an initial pose estimation using a convolutional neural network.

CONCLUSIONS: DeepIterReg demonstrates promising performance for 3D-to-2D stroke intervention image registration, potentially aiding clinicians by improving spatial understanding during EVT and reducing dependence on manual adjustments.

PMID:40407997 | DOI:10.1007/s11548-025-03412-2

Magn Reson Med. 2025 May 23. doi: 10.1002/mrm.30577. Online ahead of print.

ABSTRACT

PURPOSE: We aimed to develop a fully automatic, referenceless method for correcting distortions in echo-planar imaging (EPI) data sets, specifically designed for applications in retrospective studies lacking reference field maps or reversed-gradient scans. This work primarily targets data sets acquired with anterior-posterior or posterior-anterior phase-encoding protocols.

METHODS: Our approach used a generative adversarial network to generate a displacement map. The network model took a three-dimensional raw b0 volume from a diffusion-tensor data set as input and reproduced a displacement map, similar to that originally derived using a reversed-gradient correction method. This generative displacement map was used to correct echo-planar images across an entire diffusion data set.

RESULTS: The performance of our method was evaluated across multiple institutions using large-scale databases. We found that it effectively reduced geometric distortions in EPI data sets and improved the accuracy of diffusion indices. Moreover, it significantly enhanced the coregistration between EPI and high-resolution T1-weighted images (p < 0.01).

CONCLUSIONS: Our referenceless EPI distortion correction method has been publicly shared as a standalone application and offers a practical solution for enhancing the quality of EPI data sets in retrospective studies. It effectively reduces distortions and increases the accuracy of diffusion measures, making it a valuable tool for studies where EPI data contain no distortion calibration scan.

PMID:40407812 | DOI:10.1002/mrm.30577

Antibodies (Basel). 2025 May 14;14(2):41. doi: 10.3390/antib14020041.

ABSTRACT

BACKGROUND/OBJECTIVES: Venezuelan equine encephalitis virus (VEEV) represents a significant biothreat with no FDA-approved vaccine currently available, highlighting the need for alternative therapeutic strategies. Single-domain antibodies (sdAbs) present a potential alternative to conventional antibodies, due to their small size and ability to recognize cryptic epitopes.

METHODS: This research describes the development and preliminary evaluation of VEEV-binding sdAbs generated using a generative artificial intelligence (AI) platform. Using a dataset of known alphavirus-binding sdAbs, the AI model produced sequences with predicted affinity for the E2 glycoprotein of VEEV. These candidate sdAbs were expressed in a bacterial periplasmic system and purified for initial assessment.

RESULTS: Enzyme-linked immunosorbent assays (ELISAs) indicated binding activity of the sdAbs to VEEV antigens. In vitro neutralization tests suggested inhibition of VEEV infection in cultured cells for some of the candidates.

CONCLUSIONS: This study demonstrates how generative AI can expedite antiviral therapeutic development and establishes a framework for quick responses to emerging viral threats when extensive example databases are unavailable. Additional refinement and validation of AI-generated sdAbs could establish effective VEEV therapeutics.

PMID:40407693 | DOI:10.3390/antib14020041

Audiol Res. 2025 May 8;15(3):56. doi: 10.3390/audiolres15030056.

ABSTRACT

Background/Objectives: Cochlear implantation is an advantageous procedure for individuals with severe to profound hearing loss in many aspects related to auditory performance, social communication and quality of life. As machine learning applications have been used in the field of Otorhinolaryngology and Audiology in recent years, signal processing, speech perception and personalised optimisation of cochlear implantation are discussed. Methods: A comprehensive literature review was conducted in accordance with the PRISMA guidelines. PubMed, Scopus, Web of Science, Google Scholar and IEEE databases were searched for studies published between 2010 and 2025. We analyzed 59 articles that met the inclusion criteria. Rayyan AI software was used to classify the studies so that the risk of bias was reduced. Study design, machine learning algorithms, and audiological measurements were evaluated in the data analysis. Results: Machine learning applications were classified as preoperative evaluation, speech perception, and speech understanding in noise and other studies. The success rates of the articles are presented together with the number of articles changing over the years. It was observed that Random Forest, Decision Trees (96%), Bayesian Linear Regression (96.2%) and Extreme machine learning (99%) algorithms reached high accuracy rates. Conclusions: In cochlear implantation applications in the field of audiology, it has been observed that studies have been carried out with a variable number of people and data sets in different subfields. In machine learning applications, it is seen that a high amount of data, data diversity and long training times contribute to achieving high performance. However, more research is needed on deep learning applications in complex problems such as comprehension in noise that require time series processing. Funding and other resources: This study was not funded by any institution or organization. No registration was performed for this study.

PMID:40407670 | DOI:10.3390/audiolres15030056

Cancer Control. 2025 Jan-Dec;32:10732748251343245. doi: 10.1177/10732748251343245. Epub 2025 May 23.

ABSTRACT

IntroductionThe use of artificial intelligence (AI) in oncology has increased rapidly, transforming various healthcare areas such as pathology, radiology, diagnostics, prognosis, genomics, treatment planning, and clinical trials. However, perspectives, comfort levels, and concerns about AI in cancer care remain largely unexplored.Materials and MethodsThis prospective, descriptive cross-sectional survey study was conducted between May 20, 2024 and October 22, 2024, among 363 patients with cancer from two different hospitals affiliated with Ankara University, a tertiary care center in Türkiye. The survey included three distinct sections: (1) Perceptions: Patients' general views on AI's impact in oncology; (2) Attitudes: Comfort level with AI performing medical tasks; (3) Concerns: Specific fears related to AI implementation (eg, diagnostic errors, data privacy, healthcare costs). Survey responses were summarized descriptively, and differences by age, gender, and education were analyzed using chi-square tests.ResultsA majority (50.7%) believed AI would somewhat (32%) or significantly (18.7%) improve healthcare. However, one-third of patients (33.1%) were very uncomfortable with AI diagnosing cancer, with higher discomfort among less-educated participants (P < .005). Top patient concerns included AI making incorrect diagnoses (31.1%), increasing healthcare costs (27.5%), and not keeping data private (19.6%). Patients with higher education levels expressed less discomfort and fewer concerns.ConclusionsPatients' perceptions and attitudes on AI varied significantly based on education, highlighting the need for targeted educational initiatives. While AI holds potential to revolutionize cancer care, addressing concerns about accuracy, security, and transparency is critical to enhance its acceptance and effectiveness in clinical practice.

PMID:40407404 | DOI:10.1177/10732748251343245

Rev Sci Instrum. 2025 May 1;96(5):051501. doi: 10.1063/5.0252130.

ABSTRACT

Since the establishment of global goals for carbon neutrality and peak carbon emissions, optimizing renewable energy use has become a global priority. Wind turbine blades, as core components of wind power systems, require effective health monitoring and damage identification to ensure stable turbine operation and enhance economic efficiency. This paper applies bibliometric analysis to classify existing blade damage detection methods, comparing major non-destructive testing techniques, including strain data monitoring, vibration data monitoring, acoustic measurement, ultrasonic testing, thermal imaging, and image recognition. This paper discusses the application scenarios, strengths, and limitations of each technique, with an emphasis on future trends, and includes damage assessment through multi-method integration, advancements in online and non-destructive damage detection technologies, and the application of intelligent algorithms, such as deep learning. This study aims to guide wind power professionals in selecting blade health monitoring technologies, thereby promoting sustainability and efficiency in the wind power industry.

PMID:40407392 | DOI:10.1063/5.0252130