Deep learning

Automated Euler number of the alveolar capillary network based on deep learning segmentation with verification by stereological methods

Diseases like bronchopulmonary dysplasia (BPD) affect the development of the pulmonary vasculature, including the alveolar capillary network (ACN). Since pulmonary development is highly dependent on angiogenesis and microvascular maturation, ACN investigations are essential. Therefore, efficient methods are needed for quantitative comparative studies. Here, the suitability of deep learning (DL) for processing serial block-face scanning electron microscopic (SBF-SEM) data by generating ACN segmentations, 3D reconstructions and performing automated quantitative analyses based on them, was tested. Since previous studies revealed inefficient ACN segmentation as the limiting factor in the overall workflow, a 2D DL-based approach was used with existing data, aiming at the reduction of necessary manual interaction. Automated quantitative analyses based on completed segmentations were performed subsequently. The results were compared to stereological estimations, assessing segmentation quality and result reliability. It was shown that the DL-based approach was suitable for generating segmentations on SBF-SEM data. This approach generated more complete initial ACN segmentations than an established method, despite the limited amount of available training data and the use of a 2D rather than a 3D approach. The quality of the completed ACN segmentations was assessed as sufficient. Furthermore, quantitative analyses delivered reliable results about the ACN architecture, automatically obtained contrary to manual stereological approaches. This study demonstrated that ACN segmentation is still the part of the overall workflow that requires improvement regarding the reduction of manual interaction to benefit from available automated software tools. Nevertheless, the results indicated that it could be advantageous taking further efforts to implement a 3D DL-based segmentation approach. As the amount of analysed data was limited, this study was not conducted to obtain representative data about BPD-induced ACN alterations, but to highlight next steps towards a fully automated segmentation and evaluation workflow, enabling larger sample sizes and representative studies.

Universal photonic artificial intelligence acceleration

Over the past decade, photonics research has explored accelerated tensor operations, foundational to artificial intelligence (AI) and deep learning1-4, as a path towards enhanced energy efficiency and performance5-14. The field is centrally motivated by finding alternative technologies to extend computational progress in a post-Moore's law and Dennard scaling era15-19. Despite these advances, no photonic chip has achieved the precision necessary for practical AI applications, and demonstrations have been limited to simplified benchmark tasks. Here we introduce a photonic AI processor that executes advanced AI models, including ResNet3 and BERT20,21, along with the Atari deep reinforcement learning algorithm originally demonstrated by DeepMind22. This processor achieves near-electronic precision for many workloads, marking a notable entry for photonic computing into competition with established electronic AI accelerators23 and an essential step towards developing post-transistor computing technologies.

Generative Artificial Intelligence in Pathology and Medicine: A Deeper Dive

This review article builds upon the introductory piece in our 7-part series, delving deeper into the transformative potential of generative artificial intelligence (Gen AI) in pathology and medicine. The article explores the applications of Gen AI models in pathology and medicine, including the use of custom chatbots for diagnostic report generation, synthetic image synthesis for training new models, data set augmentation, hypothetical scenario generation for educational purposes, and the use of multimodal along with multiagent models. This article also provides an overview of the common categories within Gen AI models, discussing open-source and closed-source models, as well as specific examples of popular models such as GPT-4, Llama, Mistral, DALL-E, Stable Diffusion, and their associated frameworks (eg, transformers, generative adversarial networks, diffusion-based neural networks), along with their limitations and challenges, especially within the medical domain. We also review common libraries and tools that are currently deemed necessary to build and integrate such models. Finally, we look to the future, discussing the potential impact of Gen AI on health care, including benefits, challenges, and concerns related to privacy, bias, ethics, application programming interface costs, and security measures.

Multidimensional Directionality-Enhanced Segmentation via large vision model

Optical Coherence Tomography (OCT) facilitates a comprehensive examination of macular edema and associated lesions. Manual delineation of retinal fluid is labor-intensive and error-prone, necessitating an automated diagnostic and therapeutic planning mechanism. Conventional supervised learning models are hindered by dataset limitations, while Transformer-based large vision models exhibit challenges in medical image segmentation, particularly in detecting small, subtle lesions in OCT images. This paper introduces the Multidimensional Directionality-Enhanced Retinal Fluid Segmentation framework (MD-DERFS), which reduces the limitations inherent in conventional supervised models by adapting a transformer-based large vision model for macular edema segmentation. The proposed MD-DERFS introduces a Multi-Dimensional Feature Re-Encoder Unit (MFU) to augment the model's proficiency in recognizing specific textures and pathological features through directional prior extraction and an Edema Texture Mapping Unit (ETMU), a Cross-scale Directional Insight Network (CDIN) furnishes a holistic perspective spanning local to global details, mitigating the large vision model's deficiencies in capturing localized feature information. Additionally, the framework is augmented by a Harmonic Minutiae Segmentation Equilibrium loss (LHMSE) that can address the challenges of data imbalance and annotation scarcity in macular edema datasets. Empirical validation on the MacuScan-8k dataset shows that MD-DERFS surpasses existing segmentation methodologies, demonstrating its efficacy in adapting large vision models for boundary-sensitive medical imaging tasks. The code is publicly available at https://github.com/IMOP-lab/MD-DERFS-Pytorch.git.

Meta-Modulation: A General Learning Framework for Cross-Task Adaptation

Building learning systems possessing adaptive flexibility to different tasks is critical and challenging. In this article, we propose a novel and general meta-learning framework, called meta-modulation (MeMo), to foster the adaptation capability of a base learner across different tasks where only a few training data are available per task. For one independent task, MeMo proceeds like a "feedback regulation system," which achieves an adaptive modulation on the so-called definitive embeddings of query data to maximize the corresponding task objective. Specifically, we devise a type of efficient feedback information, definitive embedding feedback (DEF), to mathematize and quantify the unsuitability between the few training data and the base learner as well as the promising adjustment direction to reduce this unsuitability. The DEFs are encoded into high-level representation and temporarily stored as task-specific modulator templates by a modulation encoder. For coming query data, we develop an attention mechanism acting upon these modulator templates and combine both task/data-level modulation to generate the final data-specific meta-modulator. This meta-modulator is then used to modulate the query's embedding for correct decision-making. Our framework is scalable for various base learner models like multi-layer perceptron (MLP), long short-term memory (LSTM), convolutional neural network (CNN), and transformer, and applicable to different learning problems like language modeling and image recognition. Experimental results on a 2-D point synthetic dataset and various benchmarks in language and vision domains demonstrate the effectiveness and competitiveness of our framework.

In-Sensor Polarization Convolution Based on Ferroelectric-Reconfigurable Polarization-Sensitive Photodiodes

In-sensor computing can enhance the imaging system performance by putting part of the computations into the sensor. While substantial advancements have been made in latency, spectral range, and functionalities, the strategy for in-sensor light polarization computing has remained unexplored. Here, it is shown that ferroelectric-reconfigurable polarization-sensitive photodiodes (FPPDs) based on BiFeO3 nanowire arrays can perform in-sensor computations on polarization information. This innovation leverages the anisotropic photoresponse from the 1D structure of nanowires and the non-volatile reconfigurability of ferroelectrics. The devices show programmable anisotropic ratios as high as 5219, surpassing most state-of-the-art polarization-sensitive photodetectors and commercial polarization image sensors. Employing tunable photoresponse as kernel, FPPDs can perform convolutions to directly extract feature maps containing polarization information, which raises the recognition accuracy on road-scene objects under adverse weather up to 89.6%. The research highlights the potential of FPPDs as a highly efficient vision sensor and extends the boundaries of advanced intelligent imaging systems.

Revolutionizing prostate cancer therapy: Artificial intelligence - Based nanocarriers for precision diagnosis and treatment

Prostate cancer is one of the major health challenges in the world and needs novel therapeutic approaches to overcome the limitations of conventional treatment. This review delineates the transformative potential of artificial intelligence (AL) in enhancing nanocarrier-based drug delivery systems for prostate cancer therapy. With its ability to optimize nanocarrier design and predict drug delivery kinetics, AI has revolutionized personalized treatment planning in oncology. We discuss how AI can be integrated with nanotechnology to address challenges related to tumor heterogeneity, drug resistance, and systemic toxicity. Emphasis is placed on strong AI-driven advancements in the design of nanocarriers, structural optimization, targeting of ligands, and pharmacokinetics. We also give an overview of how AI can better predict toxicity, reduce costs, and enable personalized medicine. While challenges persist in the way of data accessibility, regulatory hurdles, and interactions with the immune system, future directions based on explainable AI (XAI) models, integration of multimodal data, and green nanocarrier designs promise to move the field forward. Convergence between AI and nanotechnology has been one key step toward safer, more effective, and patient-tailored cancer therapy.

Embracing a Penta helix hub framework for co-creating sustaining and potentially disruptive sterilization innovation that enables artificial intelligence and sustainability: A scoping review

The supply of safe pipeline medical devices is of paramount importance. Opportunities exist to transform reusable medical devices for improved processing that meets diverse patient needs. There is increased interest in multi-actor hub frameworks to meet innovation challenges globally. The purpose of this scoping paper was to identify critical decontamination and sterilization needs for the medtech and pharmaceutical sectors with a focus on understanding how to effectively use the Penta helix hub framework that combines academia, industry, healthcare, policy-makers/regulators and patients/society. A PRISMA scoping review of PubMed publications was conducted over the period 2010 to January 2025. Thirty of the 124 'helix hub' papers addressed innovation where only 3 of 16 healthcare-focused helices used or mentioned the need for key performance indicators (KPIs). Early-phase helix innovation ecosystems are mainly supported by qualitative or non-empirical data. This review explores multi-actor needs along with describing quantifiable KPIs at micro (end-user), meso (innovation hub) and macro (regional, national and international) levels. This integrated Penta hub approach will help to effectively plan, co-create, manage, analyse and utilize voluminous data, for example there are ca. 60,000 and 56,000 publications per year on artificial intelligence (AI) and medical devices respectively along, with some 35,000 adverse reports on devices submitted to the US FDA. This review addresses sustaining and potentially disruptive opportunities for decontamination and sterilization that includes the use of AI-enabled devices, bespoke training and sustainability.

A perspective on automated rapid eye movement sleep assessment

Rapid eye movement sleep is associated with distinct changes in various biomedical signals that can be easily captured during sleep, lending themselves to automated sleep staging using machine learning systems. Here, we provide a perspective on the critical characteristics of biomedical signals associated with rapid eye movement sleep and how they can be exploited for automated sleep assessment. We summarise key historical developments in automated sleep staging systems, having now achieved classification accuracy on par with human expert scorers and their role in the clinical setting. We also discuss rapid eye movement sleep assessment with consumer sleep trackers and its potential for unprecedented sleep assessment on a global scale. We conclude by providing a future outlook of computerised rapid eye movement sleep assessment and the role AI systems may play.

Importance of dataset design in developing robust U-Net models for label-free cell morphology evaluation

Advances in regenerative medicine highlighted the need for label-free cell image analysis to replace conventional microscopic observation for non-invasive cell quality evaluation. Image-based evaluation provides an efficient, quantitative, and automated approach to cell analysis, but segmentation remains a critical and challenging step. In this study, we investigated how training dataset design influenced the robustness of U-Net models for cell segmentation, focusing on challenges posed by limited data availability in cell culture. Using 2592 image pairs from four cell types representing key morphological categories, we constructed 42 investigation patterns to evaluate the effects of dataset size, dataset content, and morphological diversity on model performance. Our results showed that robust segmentation models could be developed with approximately 10 raw images captured using a 4× objective lens, a much smaller dataset than typically assumed. The dataset content was found to be crucial: training dataset images that captured commonly observed cell patterns yielded more robust models compared to those capturing rare or irregular cell patterns, which often impaired model performance with large deviations. Additionally, including both spindle and round cell morphologies in the training datasets improved model robustness when tested across all four cell types, while datasets restricted to a single morphology type could not achieve robust models. These findings highlight the importance of curating datasets that capture representative yet diverse cell morphologies. By addressing critical questions about dataset design, this study provides actionable guidance for the effective use of deep learning-based cell segmentation models in manufacturing and research applications.

Ethical challenges and evolving strategies in the integration of artificial intelligence into clinical practice

Artificial intelligence (AI) has rapidly transformed various sectors, including healthcare, where it holds the potential to transform clinical practice and improve patient outcomes. However, its integration into medical settings brings significant ethical challenges that need careful consideration. This paper examines the current state of AI in healthcare, focusing on five critical ethical concerns: justice and fairness, transparency, patient consent and confidentiality, accountability, and patient-centered and equitable care. These concerns are particularly pressing as AI systems can perpetuate or even exacerbate existing biases, often resulting from non-representative datasets and opaque model development processes. The paper explores how bias, lack of transparency, and challenges in maintaining patient trust can undermine the effectiveness and fairness of AI applications in healthcare. In addition, we review existing frameworks for the regulation and deployment of AI, identifying gaps that limit the widespread adoption of these systems in a just and equitable manner. Our analysis provides recommendations to address these ethical challenges, emphasizing the need for fairness in algorithm design, transparency in model decision-making, and patient-centered approaches to consent and data privacy. By highlighting the importance of continuous ethical scrutiny and collaboration between AI developers, clinicians, and ethicists, we outline pathways for achieving more responsible and inclusive AI implementation in healthcare. These strategies, if adopted, could enhance both the clinical value of AI and the trustworthiness of AI systems among patients and healthcare professionals, ensuring that these technologies serve all populations equitably.

Benchmarking deep learning for automated peak detection on GIWAXS data

Recent advancements in X-ray sources and detectors have dramatically increased data generation, leading to a greater demand for automated data processing. This is particularly relevant for real-time grazing-incidence wide-angle X-ray scattering (GIWAXS) experiments which can produce hundreds of thousands of diffraction images in a single day at a synchrotron beamline. Deep learning (DL)-based peak-detection techniques are becoming prominent in this field, but rigorous benchmarking is essential to evaluate their reliability, identify potential problems, explore avenues for improvement and build confidence among researchers for seamless integration into their workflows. However, the systematic evaluation of these techniques has been hampered by the lack of annotated GIWAXS datasets, standardized metrics and baseline models. To address these challenges, we introduce a comprehensive framework comprising an annotated experimental dataset, physics-informed metrics adapted to the GIWAXS geometry and a competitive baseline - a classical, non-DL peak-detection algorithm optimized on our dataset. Furthermore, we apply our framework to benchmark a recent DL solution trained on simulated data and discover its superior performance compared with our baseline. This analysis not only highlights the effectiveness of DL methods for identifying diffraction peaks but also provides insights for further development of these solutions.

Optimization of sparse-view CT reconstruction based on convolutional neural network

BACKGROUND

Sparse-view CT shortens scan time and reduces radiation dose but results in severe streak artifacts due to insufficient sampling data. Deep learning methods can now suppress these artifacts and improve image quality in sparse-view CT reconstruction.

PURPOSE

The quality of sparse-view CT reconstructed images can still be improved. Additionally, the interpretability of deep learning-based optimization methods for these reconstruction images is lacking, and the role of different network layers in artifact removal requires further study. Moreover, the optimization capability of these methods for reconstruction images from various sparse views needs enhancement. This study aims to improve the network's optimization ability for sparse-view reconstructed images, enhance interpretability, and boost generalization by establishing multiple network structures and datasets.

METHODS

In this paper, we developed a sparse-view CT reconstruction images improvement network (SRII-Net) based on U-Net. We added a copy pathway in the network and designed a residual image output block to boost the network's performance. Multiple networks with different connectivity structures were established using SRII-Net to analyze the contribution of each layer to artifact removal, improving the network's interpretability. Additionally, we created multiple datasets with reconstructed images of various sampling views to train and test the proposed network, investigating how these datasets from different sampling views affect the network's generalization ability.

RESULTS

The results show that the proposed method outperforms current networks, with significant improvements in metrics like PSNR and SSIM. Image optimization time is at the millisecond level. By comparing the performance of different network structures, we've identified the impact of various hierarchical structures. The image detail information learned by shallow layers and the high-level abstract feature information learned by deep layers play a crucial role in optimizing sparse-view CT reconstruction images. Training the network with multiple mixed datasets revealed that, under a certain amount of data, selecting the appropriate categories of sampling views and their corresponding samples can effectively enhance the network's optimization ability for reconstructing images with different sampling views.

CONCLUSIONS

The network in this paper effectively suppresses artifacts in reconstructed images with different sparse views, improving generalization. We have also created diverse network structures and datasets to deepen the understanding of artifact removal in deep learning networks, offering insights for noise reduction and image enhancement in other imaging methods.

Making sense of transformer success

This article provides an epistemological analysis of current attempts of explaining how the relatively simple algorithmic components of neural language models (NLMs) provide them with genuine linguistic competence. After introducing the Transformer architecture, at the basis of most of current NLMs, the paper firstly emphasizes how the central question in the philosophy of AI has been shifted from "can machines think?", as originally put by Alan Turing, to "how can machines think?", pointing to an explanatory gap for NLMs. Subsequently, existing explanatory strategies for the functioning of NLMs are analyzed to argue that they, however debated, do not differ from the explanatory strategies used in cognitive science to explain intelligent behaviors of humans. In particular, available experimental studies turned to test the theory of mind, discourse entity tracking, and property induction in NLMs are examined under the light of the functional analysis in the philosophy of cognitive science; the so-called copying algorithm and the induction head phenomenon of a Transformer are shown to provide a mechanist explanation of in-context learning; finally, current pioneering attempts to use NLMs to predict brain activation patterns when processing language are here shown to involve what we call a co-simulation, in which a NLM and the brain are used to simulate and understand each other.

Automatic detection for bioacoustic research: a practical guide from and for biologists and computer scientists

Recent years have seen a dramatic rise in the use of passive acoustic monitoring (PAM) for biological and ecological applications, and a corresponding increase in the volume of data generated. However, data sets are often becoming so sizable that analysing them manually is increasingly burdensome and unrealistic. Fortunately, we have also seen a corresponding rise in computing power and the capability of machine learning algorithms, which offer the possibility of performing some of the analysis required for PAM automatically. Nonetheless, the field of automatic detection of acoustic events is still in its infancy in biology and ecology. In this review, we examine the trends in bioacoustic PAM applications, and their implications for the burgeoning amount of data that needs to be analysed. We explore the different methods of machine learning and other tools for scanning, analysing, and extracting acoustic events automatically from large volumes of recordings. We then provide a step-by-step practical guide for using automatic detection in bioacoustics. One of the biggest challenges for the greater use of automatic detection in bioacoustics is that there is often a gulf in expertise between the biological sciences and the field of machine learning and computer science. Therefore, this review first presents an overview of the requirements for automatic detection in bioacoustics, intended to familiarise those from a computer science background with the needs of the bioacoustics community, followed by an introduction to the key elements of machine learning and artificial intelligence that a biologist needs to understand to incorporate automatic detection into their research. We then provide a practical guide to building an automatic detection pipeline for bioacoustic data, and conclude with a discussion of possible future directions in this field.

Efficient diagnosis of retinal disorders using dual-branch semi-supervised learning (DB-SSL): An enhanced multi-class classification approach

The early diagnosis of retinal disorders is essential in preventing permanent or partial blindness. Identifying these conditions promptly guarantees early treatment and prevents blindness. However, the challenge lies in accurately diagnosing these conditions, especially with limited labeled data. This study aims to enhance the diagnostic accuracy of retinal disorders using a novel Dual-Branch Semi-Supervised Learning (DB-SSL) approach that leverages both labeled and unlabeled data for multi-class classification of eye diseases. Employing Color Fundus Photography (CFP), our research integrates a Convolutional Neural Network (CNN) that integrates features from two parallel branches. This framework effectively handles the complexity of ocular imaging by utilizing self-training-based semi-supervised learning to explore relationships within unlabeled data. We propose and evaluate six CNN models: ResNet50, DenseNet121, MobileNetV2, EfficientNetB0, SqueezeNet1_0, and a hybrid of ResNet50 and MobileNetV2 on their ability to classify four key eye conditions: cataract, diabetic retinopathy, glaucoma, and normal, using a large, diverse OIH dataset containing 4217 fundus images. Among the evaluated models, ResNet50 emerged as the most accurate, achieving 93.14 % accuracy on unseen data. The model demonstrates robustness with a sensitivity of 93 % and specificity of 98.37 %, along with a precision and F1 Score of 93 % each, and a Cohen's Kappa of 90.85 %. Additionally, it exhibits an AUC score of 97.75 % nearing perfection. Systematically removing certain components from the ResNet50 model further validates its efficacy. Our findings underscore the potential of advanced CNN architectures combined with semi-supervised learning in enhancing the accuracy of eye disease classification systems, particularly in resource-constrained environments where the procurement of large labeled datasets is challenging and expensive. This approach is well-suited for integration into Clinical Decision Support Systems (CDSS), providing valuable diagnostic assistance in real-world clinical settings.

Cooking loss estimation of semispinalis capitis muscle of pork butt using a deep neural network on hyperspectral data

This study evaluated the performance of a deep-learning-based model that predicted cooking loss in the semispinalis capitis (SC) muscle of pork butts using hyperspectral images captured 24 h postmortem. To overcome low-scale samples, 70 pork butts were used with pixel-based data augmentation. Principal component regression (PCR) and partial least squares regression (PLSR) models for predicting cooking loss in SC muscle showed higher R2 values with multiplicative signal correction, while the first derivative resulted in a lower root mean square error (RMSE). The deep learning-based model outperformed the PCR and PLSR models. The classification accuracy of the models for cooking loss grade classification decreased as the number of grades increased, with the models with three grades achieving the highest classification accuracy. The deep learning model exhibited the highest classification accuracy (0.82). Cooking loss in the SC muscle was visualized using a deep learning model. The pH and cooking loss of the SC muscle were significantly correlated with the cooking loss of pork butt slices (-0.54 and 0.69, respectively). Therefore, a deep learning model using hyperspectral images can predict the cooking loss grade of SC muscle. This suggests that nondestructive prediction of the quality properties of pork butts can be achieved using hyperspectral images obtained from the SC muscle.

Utilizing machine learning for predicting drug release from polymeric drug delivery systems

Polymeric drug delivery systems (PDDS) play a crucial role in controlled drug release, providing improved therapeutic outcomes. However, formulating PDDS and predicting their release profiles remain challenging due to their complex structures and the numerous variables that influence their behavior. Traditional mathematical and empirical prediction methods are limited in capturing these complexities. Recent studies have unveiled the potential of Machine Learning (ML) in revolutionizing drug delivery, particularly in formulating complex PDDS. This article provides an overview of the significant and fundamental principles of various ML strategies in estimating PDDS drug release behavior. Our focus extends to the accomplishments and pivotal discoveries in current research, spanning seven distinct sustained-release drug delivery systems: matrix tablets, microspheres, implants, hydrogels, films, 3D-printed dosage forms, and other innovations. Furthermore, it addresses the challenges associated with ML-based drug release prediction and presents current solutions while delving into future perspectives. Our investigation underscores the significance of Artificial Neural Networks in ML-based PDDS release profile prediction, surpassing both traditional and alternative ML-based methods. These extensive datasets can be drawn from literature-based resources or enhanced through specific algorithms. Moreover, ensemble-based models have proven advantageous in scenarios involving intricate relationships, such as a high number of output parameters. ML-based drug release prediction notably exhibits substantial promise in 3D-printed dosage forms, presenting a frontier for personalized medicine and precise drug delivery.

Artificial intelligence in dentistry: Exploring emerging applications and future prospects

OBJECTIVES

This narrative review aimed to explore the evolution and advancements of artificial intelligence technologies, highlighting their transformative impact on healthcare, education, and specific aspects within dentistry as a field.

DATA AND SOURCES

Subtopics within artificial intelligence technologies in dentistry were identified and divided among four reviewers. Electronic searches were performed with search terms that included: artificial intelligence, technologies, healthcare, education, dentistry, restorative, prosthodontics, periodontics, endodontics, oral surgery, oral pathology, oral medicine, implant dentistry, dental education, dental patient care, dental practice management, and combinations of the keywords.

STUDY

selection: A total of 120 articles were included for review that evaluated the use of artificial intelligence technologies within the healthcare and dental field. No formal evidence-based quality assessment was performed due to the narrative nature of this review. The conducted search was limited to the English language with no other further restrictions.

RESULTS

The significance and applications of artificial intelligence technologies on the areas of dental education, dental patient care, and dental practice management were reviewed, along with the existing limitations and future directions of artificial intelligence in dentistry as whole. Current artificial intelligence technologies have shown promising efforts to bridge the gap between theoretical knowledge and clinical practice in dental education, as well as improved diagnostic information gathering and clinical decision-making abilities in patient care throughout various dental specialties. The integration of artificial intelligence into patient administration aspects have enabled practices to develop more efficient management workflows.

CONCLUSIONS

Despite the limitations that exist, the integration of artificial intelligence into the dental profession comes with numerous benefits that will continue to evolve each day. While the challenges and ethical considerations, mainly concerns about data privacy, are areas that need to be further addressed, the future of artificial intelligence in dentistry looks promising, with ongoing research aimed at overcoming current limitations and expanding artificial intelligence technologies.

Interpretable discovery of patterns in tabular data via spatially semantic topographic maps

Tabular data-rows of samples and columns of sample features-are ubiquitously used across disciplines. Yet the tabular representation makes it difficult to discover underlying associations in the data and thus hinders their analysis and the discovery of useful patterns. Here we report a broadly applicable strategy for unravelling intertwined relationships in tabular data by reconfiguring each data sample into a spatially semantic 2D topographic map, which we refer to as TabMap. A TabMap preserves the original feature values as pixel intensities, with the relationships among the features spatially encoded in the map (the strength of two inter-related features correlates with their distance on the map). TabMap makes it possible to apply 2D convolutional neural networks to extract association patterns in the data to aid data analysis, and offers interpretability by ranking features according to importance. We show the superior predictive performance of TabMap by applying it to 12 datasets across a wide range of biomedical applications, including disease diagnosis, human activity recognition, microbial identification and the analysis of quantitative structure-activity relationships.

Larger sample sizes are needed when developing a clinical prediction model using machine learning in oncology: methodological systematic review

BACKGROUND AND OBJECTIVES

Having a sufficient sample size is crucial when developing a clinical prediction model. We reviewed details of sample size in studies developing prediction models for binary outcomes using machine learning (ML) methods within oncology and compared the sample size used to develop the models with the minimum required sample size needed when developing a regression-based model (Nmin).

METHODS

We searched the Medline (via OVID) database for studies developing a prediction model using ML methods published in December 2022. We reviewed how sample size was justified. We calculated Nmin, which is the Nmin, and compared this with the sample size that was used to develop the models.

RESULTS

Only one of 36 included studies justified their sample size. We were able to calculate Nmin for 17 (47%) studies. 5/17 studies met Nmin, allowing to precisely estimate the overall risk and minimize overfitting. There was a median deficit of 302 participants with the event (n = 17; range: -21,331 to 2298) when developing the ML models. An additional three out of the 17 studies met the required sample size to precisely estimate the overall risk only.

CONCLUSION

Studies developing a prediction model using ML in oncology seldom justified their sample size and sample sizes were often smaller than Nmin. As ML models almost certainly require a larger sample size than regression models, the deficit is likely larger. We recommend that researchers consider and report their sample size and at least meet the minimum sample size required when developing a regression-based model.

An integrated large-scale photonic accelerator with ultralow latency

Integrated photonics, particularly silicon photonics, have emerged as cutting-edge technology driven by promising applications such as short-reach communications, autonomous driving, biosensing and photonic computing1-4. As advances in AI lead to growing computing demands, photonic computing has gained considerable attention as an appealing candidate. Nonetheless, there are substantial technical challenges in the scaling up of integrated photonics systems to realize these advantages, such as ensuring consistent performance gains in upscaled integrated device clusters, establishing standard designs and verification processes for complex circuits, as well as packaging large-scale systems. These obstacles arise primarily because of the relative immaturity of integrated photonics manufacturing and the scarcity of advanced packaging solutions involving photonics. Here we report a large-scale integrated photonic accelerator comprising more than 16,000 photonic components. The accelerator is designed to deliver standard linear matrix multiply-accumulate (MAC) functions, enabling computing with high speed up to 1 GHz frequency and low latency as small as 3 ns per cycle. Logic, memory and control functions that support photonic matrix MAC operations were designed into a cointegrated electronics chip. To seamlessly integrate the electronics and photonics chips at the commercial scale, we have made use of an innovative 2.5D hybrid advanced packaging approach. Through the development of this accelerator system, we demonstrate an ultralow computation latency for heuristic solvers of computationally hard Ising problems whose performance greatly relies on the computing latency.

Systems Human Immunology and AI: Immune Setpoint and Immune Health

The immune system, critical for human health and implicated in many diseases, defends against pathogens, monitors physiological stress, and maintains tissue and organismal homeostasis. It exhibits substantial variability both within and across individuals and populations. Recent technological and conceptual progress in systems human immunology has provided predictive insights that link personal immune states to intervention responses and disease susceptibilities. Artificial intelligence (AI), particularly machine learning (ML), has emerged as a powerful tool for analyzing complex immune data sets, revealing hidden patterns across biological scales, and enabling predictive models for individualistic immune responses and potentially personalized interventions. This review highlights recent advances in deciphering human immune variation and predicting outcomes, particularly through the concepts of immune setpoint, immune health, and use of the immune system as a window for measuring health. We also provide a brief history of AI; review ML modeling approaches, including their applications in systems human immunology; and explore the potential of AI to develop predictive models and personal immune state embeddings to detect early signs of disease, forecast responses to interventions, and guide personalized health strategies.

An Interpretable Adaptive Multiscale Attention Deep Neural Network for Tabular Data

Deep learning (DL) has been demonstrated to be a valuable tool for analyzing signals such as sounds and images, thanks to its capabilities of automatically extracting relevant patterns as well as its end-to-end training properties. When applied to tabular structured data, DL has exhibited some performance limitations compared to shallow learning techniques. This work presents a novel technique for tabular data called adaptive multiscale attention deep neural network architecture (also named excited attention). By exploiting parallel multilevel feature weighting, the adaptive multiscale attention can successfully learn the feature attention and thus achieve high levels of F1-score on seven different classification tasks (on small, medium, large, and very large datasets) and low mean absolute errors on four regression tasks of different size. In addition, adaptive multiscale attention provides four levels of explainability (i.e., comprehension of its learning process and therefore of its outcomes): 1) calculates attention weights to determine which layers are most important for given classes; 2) shows each feature's attention across all instances; 3) understands learned feature attention for each class to explore feature attention and behavior for specific classes; and 4) finds nonlinear correlations between co-behaving features to reduce dataset dimensionality and improve interpretability. These interpretability levels, in turn, allow for employing adaptive multiscale attention as a useful tool for feature ranking and feature selection.

Generating accurate sex estimation from hand X-ray images using AI deep-learning techniques: A study of limited bone regions

Hand bone structure provides valuable features for sex estimation. This research introduces a novel approach using Artificial Intelligence (AI), specifically Convolutional Neural Networks (CNNs), to classify sex from hand X-ray images, focusing on the diagnostic potential of specific bone regions. We assess CNN performance on different hand skeleton areas, utilize Score-CAM to understand sex-discriminating features, and evaluate advanced CNN architectures. While the Xception model achieved the highest overall accuracy of 83.5% using complete hand X-rays, the InceptionResNetV2 model demonstrated remarkable efficiency by achieving 81.68% accuracy using only the proximal phalanx and metacarpal bones, maintaining a comparable AUC-ROC score of 0.92. Metacarpals of the first and second fingers were identified as key for differentiation. This approach demonstrates the power of AI in skeletal analysis and represents a significant step towards deployable AI tools for forensic and medical sex identification.

Application of Artificial Intelligence in Retinopathy of Prematurity From 2010 to 2023: A Bibliometric Analysis

Background and Aims

Retinopathy of prematurity (ROP) remains a leading cause of childhood blindness worldwide. In recent years, artificial intelligence (AI) has emerged as a powerful tool for the screening and management of ROP. This study aimed to investigate the evolving and longitudinal publication patterns related to AI in ROP using bibliometric methodologies.

Methods

We conducted a descriptive analysis of AI in ROP documents retrieved from the Web of Science database up to September 10, 2023. Data analysis and visualization were performed using Bibliometrix and VOSviewer, covering publications, journals, authors, institutions, countries, collaboration networks, keywords, and trending topics.

Results

Our analysis of 188 publications on AI in ROP revealed an average of 7.62 authors per document and a notable increase in annual publications since 2017. The United States (98/188), Oregon Health & Science University (66/188), Investigative Ophthalmology & Visual Science (29/188) and author Michael F. Chiang (60/188) led contributions. A prominent 21-country network emerged as the largest in country-level coauthorship. Key technical terms included "artificial intelligence," "deep learning," "machine learning," and "telemedicine," with a recent shift from "feature selection" to "deep learning," "machine learning" and "fundus images" in trending topics.

Conclusion

Our bibliometric analysis highlights advancements in AI research on ROP, focusing on key publication characteristics, major contributors, and emerging trends. The findings indicate that AI in ROP is a rapidly growing field. Future studies should focus on addressing the clinical implementation and ethical concerns of AI in ROP.

Atypical Developmental Patterns of Sensorimotor-Related Networks in Autism Spectrum Disorder: A BrainAGE Study Based on Resting-State fMRI

Autism spectrum disorder (ASD) is a type of neurodevelopmental disorder characterized by atypical brain development. Previous whole-brain BrainAGE studies have unveiled the presence of accelerated or delayed brain function developmental patterns in individuals with ASD. However, it remains unclear whether these patterns manifest at a global level throughout the entire brain or are specific to certain functional sub-networks. The study included resting-state functional magnetic resonance imaging (fMRI) data from 127 individuals with ASD and 135 healthy controls (aged between 5 and 40 years). ALFF maps were measured for each participant. Then, sub-network-level BrainAGE analyses were conducted across 10 sub-networks using the Individual-weighted Multilayer Perceptron Network (ILWMLP) regression method. The BrainAGE analyses revealed atypical developmental trajectories in sensorimotor-related sub-networks, encompassing auditory, motor, and sensorimotor sub-networks. In individuals with ASD, delayed brain function development was observed in the auditory and sensorimotor networks, with a more pronounced delay observed in older individuals. Conversely, the motor network exhibited accelerated development in younger individuals but delayed development in older individuals. Our findings unveiled aberrant developmental patterns in sensorimotor-related sub-networks among individuals with ASD, exhibiting distinct atypical profiles across different sub-networks. These results might contribute to a deeper understanding of the deviant brain development observed in ASD.

Microbial Transcription Factor-Based Biosensors: Innovations from Design to Applications in Synthetic Biology

Transcription factor-based biosensors (TFBs) are powerful tools in microbial biosensor applications, enabling dynamic control of metabolic pathways, real-time monitoring of intracellular metabolites, and high-throughput screening (HTS) for strain engineering. These systems use transcription factors (TFs) to convert metabolite concentrations into quantifiable outputs, enabling precise regulation of metabolic fluxes and biosynthetic efficiency in microbial cell factories. Recent advancements in TFB, including improved sensitivity, specificity, and dynamic range, have broadened their applications in synthetic biology and industrial biotechnology. Computational tools such as Cello have further revolutionized TFB design, enabling in silico optimization and construction of complex genetic circuits for integrating multiple signals and achieving precise gene regulation. This review explores innovations in TFB systems for microbial biosensors, their role in metabolic engineering and adaptive evolution, and their future integration with artificial intelligence and advanced screening technologies to overcome critical challenges in synthetic biology and industrial bioproduction.

Deep learning-based breast MRI for predicting axillary lymph node metastasis: a systematic review and meta-analysis

BACKGROUND

To perform a systematic review and meta-analysis that assesses the diagnostic performance of deep learning algorithms applied to breast MRI for predicting axillary lymph nodes metastases in patients of breast cancer.

METHODS

A systematic literature search in PubMed, MEDLINE, and Embase databases for articles published from January 2004 to February 2025. Inclusion criteria were: patients with breast cancer; deep learning using MRI images was applied to predict axillary lymph nodes metastases; sufficient data were present; original research articles. Quality Assessment of Diagnostic Accuracy Studies-AI and Checklist for Artificial Intelligence in Medical Imaging was used to assess the quality. Statistical analysis included pooling of diagnostic accuracy and investigating between-study heterogeneity. A summary receiver operating characteristic curve (SROC) was performed. R statistical software (version 4.4.0) was used for statistical analyses.

RESULTS

A total of 10 studies were included. The pooled sensitivity and specificity were 0.76 (95% CI, 0.67-0.83) and 0.81 (95% CI, 0.74-0.87), respectively, with both measures having moderate between-study heterogeneity (I2 = 61% and 60%, respectively; p < 0.01). The SROC analysis yielded a weighted AUC of 0.788.

CONCLUSION

This meta-analysis demonstrates that deep learning algorithms applied to breast MRI offer promising diagnostic performance for predicting axillary lymph node metastases in breast cancer patients. Incorporating deep learning into clinical practice may enhance decision-making by providing a non-invasive method to more accurately predict lymph node involvement, potentially reducing unnecessary surgeries.

Identification of MAD2L1 as a novel biomarker for hepatoblastoma through bioinformatics and machine learning approaches

Objective

This study aims to identify potential biomarkers for Hepatoblastoma (HB) using bioinformatics and machine learning, and to explore their underlying mechanisms of action.

Methods

We analyzed the datasets GSE131329 and GSE133039 to perform differential gene expression analysis. Single-sample gene set enrichment analysis (ssGSEA) and weighted gene co-expression network analysis (WGCNA) were utilized to identify gene modules linked to gene set activity. Protein-protein interaction (PPI) networks were constructed to identify hub genes, while random forest and support vector machine models were employed to screen for key diagnostic genes. Survival and immune infiltration analyses were conducted to assess the prognostic significance of these genes. Additionally, the expression levels, biological functions, and mechanisms of action of the selected genes were validated in HB cells through relevant experimental assays.

Results

We identified 1,377 and 1,216 differentially expressed genes in datasets GSE131329 and GSE133039, respectively. ssGSEA and WGCNA analyses identified 234 genes significantly linked to gene set activity. PPI analysis identified 20 core Hub genes. Machine learning highlighted three key diagnostic genes: CDK1, CCNA2, and MAD2L1. Studies have demonstrated that MAD2L1 is significantly overexpressed in HB and is associated with prognosis. WGCNA revealed that MAD2L1 is enriched in gene sets related to E2F_ TARGETS and G2M_CHECKPOINT. Experimental assays demonstrated that MAD2L1 knockdown significantly inhibits the proliferation, migration, and invasion of HB cell lines, and that MAD2L1 promotes cell cycle progression through the regulation of E2F.

Conclusion

Our study identifies MAD2L1 as a novel potential biomarker for HB, providing new strategies for early diagnosis and targeted therapy in HB.

Nuclear phase retrieval spectroscopy using resonant x-ray scattering

Light-matter interaction is exploited in spectroscopic techniques to access information about molecular, atomic or nuclear constituents of a sample. While scattered light carries both amplitude and phase information of the electromagnetic field, the latter is lost in intensity measurements. However, often the phase information is paramount to reconstruct the desired information of the target, as it is well known from coherent x-ray imaging. Here we introduce a phase retrieval method which allows us to reconstruct the field phase information from two-dimensional time- and energy-resolved spectra. We apply this method to the case of x-ray scattering off Mössbauer nuclei at a synchrotron radiation source. Knowledge of the phase allows also for the reconstruction of energy spectra from two-dimensional experimental data sets with excellent precision, without theoretical modelling of the sample. Our approach provides an efficient and accurate data analysis tool which will benefit x-ray quantum optics and Mössbauer spectroscopy with synchrotron radiation alike.

Diagnostic Accuracy of Deep Learning for Intracranial Hemorrhage Detection in Non-Contrast Brain CT Scans: A Systematic Review and Meta-Analysis

Background: Intracranial hemorrhage (ICH) is a life-threatening medical condition that needs early detection and treatment. In this systematic review and meta-analysis, we aimed to update our knowledge of the performance of deep learning (DL) models in detecting ICH on non-contrast computed tomography (NCCT). Methods: The study protocol was registered with PROSPERO (CRD420250654071). PubMed/MEDLINE and Google Scholar databases and the reference section of included studies were searched for eligible studies. The risk of bias in the included studies was assessed using the QUADAS-2 tool. Required data was collected to calculate pooled sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) with the corresponding 95% CI using the random effects model. Results: Seventy-three studies were included in our qualitative synthesis, and fifty-eight studies were selected for our meta-analysis. A pooled sensitivity of 0.92 (95% CI 0.90-0.94) and a pooled specificity of 0.94 (95% CI 0.92-0.95) were achieved. Pooled PPV was 0.84 (95% CI 0.78-0.89) and pooled NPV was 0.97 (95% CI 0.96-0.98). A bivariate model showed a pooled AUC of 0.96 (95% CI 0.95-0.97). Conclusions: This meta-analysis demonstrates that DL performs well in detecting ICH from NCCTs, highlighting a promising potential for the use of AI tools in various practice settings. More prospective studies are needed to confirm the potential clinical benefit of implementing DL-based tools and reveal the limitations of such tools for automated ICH detection and their impact on clinical workflow and outcomes of patients.

Advances in machine learning for keratoconus diagnosis

PURPOSE

To review studies reporting the role of Machine Learning (ML) techniques in the diagnosis of keratoconus (KC) over the past decade, shedding light on recent developments while also highlighting the existing gaps between academic research and practical implementation in clinical settings.

METHODS

The review process begins with a systematic search of primary digital libraries using relevant keywords. A rigorous set of inclusion and exclusion criteria is then applied, resulting in the identification of 62 articles for analysis. Key research questions are formulated to address advancements in ML for KC diagnosis, corneal imaging modalities, types of datasets utilised, and the spectrum of KC conditions investigated over the past decade. A significant gap between academic research and practical implementation in clinical settings is identified, forming the basis for actionable recommendations tailored for both ML developers and ophthalmologists. Additionally, a proposed roadmap model is presented to facilitate the integration of ML models into clinical practice, enhancing diagnostic accuracy and patient care.

RESULTS

The analysis revealed that the diagnosis of KC predominantly relies on supervised classifiers (97%), with Random Forest being the most used algorithm (27%), followed by Deep Learning including Convolution Neural Networks (16%), Feedforward and Feedback Neural Networks (12%), and Support Vector Machines (12%). Pentacam is identified as the leading corneal imaging modality (56%), and a substantial majority of studies (91%) utilize local datasets, primarily consisting of numerical corneal parameters (77%). The most studied KC conditions were non-KC (NKC) vs. clinical KC (CKC) (29%), NKC vs. Subclinical KC (SCKC) (24%), NKC vs. SCKC vs. CKC (20%), SCKC vs. CKC (7%). However, only 20% of studies focused on addressing KC severity stages, emphasizing the need for more research in this area. These findings highlight the current landscape of ML in KC diagnosis and uncover existing challenges, and suggest potential avenues for further research and development, with particular emphasis on the dominance of certain algorithms and imaging modalities.

CONCLUSION

Key obstacles include the lack of consensus on an objective diagnostic standard for early KC detection and severity staging, limited multidisciplinary collaboration, and restricted access to public datasets. Further research is crucial to overcome these challenges and apply findings in clinical practice.

Near-term prediction of sustained ventricular arrhythmias applying artificial intelligence to single-lead ambulatory electrocardiogram

BACKGROUND AND AIMS

Accurate near-term prediction of life-threatening ventricular arrhythmias would enable pre-emptive actions to prevent sudden cardiac arrest/death. A deep learning-enabled single-lead ambulatory electrocardiogram (ECG) may identify an ECG profile of individuals at imminent risk of sustained ventricular tachycardia (VT).

METHODS

This retrospective study included 247 254, 14 day ambulatory ECG recordings from six countries. The first 24 h were used to identify patients likely to experience sustained VT occurrence (primary outcome) in the subsequent 13 days using a deep learning-based model. The development set consisted of 183 177 recordings. Performance was evaluated using internal (n = 43 580) and external (n = 20 497) validation data sets. Saliency mapping visualized features influencing the model's risk predictions.

RESULTS

Among all recordings, 1104 (.5%) had sustained ventricular arrhythmias. In both the internal and external validation sets, the model achieved an area under the receiver operating characteristic curve of .957 [95% confidence interval (CI) .943-.971] and .948 (95% CI .926-.967). For a specificity fixed at 97.0%, the sensitivity reached 70.6% and 66.1% in the internal and external validation sets, respectively. The model accurately predicted future VT occurrence of recordings with rapid sustained VT (≥180 b.p.m.) in 80.7% and 81.1%, respectively, and 90.0% of VT that degenerated into ventricular fibrillation. Saliency maps suggested the role of premature ventricular complex burden and early depolarization time as predictors for VT.

CONCLUSIONS

A novel deep learning model utilizing dynamic single-lead ambulatory ECGs accurately identifies patients at near-term risk of ventricular arrhythmias. It also uncovers an early depolarization pattern as a potential determinant of ventricular arrhythmias events.

A review of denoising methods in single-particle cryo-EM

Cryo-EM has become a vital technique for resolving macromolecular structures at near-atomic resolution, enabling the visualization of proteins and large molecular complexes. However, the images are often accompanied by extremely low SNR, which poses significant challenges for subsequent processes such as particle picking and 3D reconstruction. Effective denoising methods can substantially improve SNR, making downstream analyzes more accurate and reliable. Thus, image denoising is an essential step in cryo-EM data processing. This paper comprehensively reviews recent advances in image denoising methods for single-particle analysis, covering approaches from traditional filtering methods to the latest deep learning-based strategies. By analyzing and comparing mainstream denoising methods, this review aims to advance the field of single-particle cryo-EM denoising, facilitate the acquisition of higher-quality images, and offer valuable insights for researchers new to the field.

Artificial Intelligence in Gas Sensing: A Review

The role of artificial intelligence (AI), machine learning (ML), and deep learning (DL) in enhancing and automating gas sensing methods and the implications of these technologies for emergent gas sensor systems is reviewed. Applications of AI-based intelligent gas sensors include environmental monitoring, industrial safety, remote sensing, and medical diagnostics. AI, ML, and DL methods can process and interpret complex sensor data, allowing for improved accuracy, sensitivity, and selectivity, enabling rapid gas detection and quantitative concentration measurements based on sophisticated multiband, multispecies sensor systems. These methods can discern subtle patterns in sensor signals, allowing sensors to readily distinguish between gases with similar sensor signatures, enabling adaptable, cross-sensitive sensor systems for multigas detection under various environmental conditions. Integrating AI in gas sensor technology represents a paradigm shift, enabling sensors to achieve unprecedented performance, selectivity, and adaptability. This review describes gas sensor technologies and AI while highlighting approaches to AI-sensor integration.

Neural dynamics of semantic control underlying generative storytelling

Storytelling has been pivotal for the transmission of knowledge across human history, yet the role of semantic control and its associated neural dynamics has been poorly investigated. Here, human participants generated stories that were either appropriate (ordinary), novel (random), or balanced (creative), while recording functional magnetic resonance imaging (fMRI). Deep language models confirmed participants adherence to task instructions. At the neural level, linguistic and visual areas exhibited neural synchrony across participants regardless of the semantic control level, with parietal and frontal regions being more synchronized during random ideation. Importantly, creative stories were differentiated by a multivariate pattern of neural activity in frontal and fronto-temporo-parietal cortices compared to ordinary and random stories. Crucially, similar brain regions were also encoding the features that distinguished the stories. Moreover, we found specific spatial frequency patterns underlying the modulation of semantic control during story generation, while functional coupling in default, salience, and control networks differentiated creative stories with their controls. Remarkably, the temporal irreversibility between visual and high-level areas was higher during creative ideation, suggesting the enhanced hierarchical structure of causal interactions as a neural signature of creative storytelling. Together, our findings highlight the neural mechanisms underlying the regulation of semantic exploration during narrative ideation.

Blind HDR image quality assessment based on aggregating perception and inference features

High Dynamic Range (HDR) images, with their expanded range of brightness and color, provide a far more realistic and immersive viewing experience compared to Low Dynamic Range (LDR) images. However, the significant increase in peak luminance and contrast inherent in HDR images often accentuates artifacts, thus limiting the effectiveness of traditional LDR-based image quality assessment (IQA) algorithms when applied to HDR content. To address this, we propose a novel blind IQA method tailored specifically for HDR images, which incorporates both the perception and inference processes of the human visual system (HVS). Our approach begins with multi-scale Retinex decomposition to generate reflectance maps with varying sensitivity, followed by the calculation of gradient similarities from these maps to model the perception process. Deep feature maps are then extracted from the last pooling layer of a pretrained VGG16 network to capture inference characteristics. These gradient similarity maps and deep feature maps are subsequently aggregated for quality prediction using support vector regression (SVR). Experimental results demonstrate that the proposed method achieves outstanding performance, outperforming other state-of-the-art HDR IQA metrics.

Exploration of geo-spatial data and machine learning algorithms for robust wildfire occurrence prediction

Wildfires play a pivotal role in environmental processes and the sustainable development of ecosystems. Timely responses can significantly reduce the damages and consequences caused by their spread. Several critical issues in wildfire behavior analysis include fire occurrence forecasting, early detection, and spread prediction. In this study, we focus on wildfire occurrence forecasting, which is a valuable tool for facilitating earlier intervention. Conventional approaches primarily rely on the computation of fire indices based on weather conditions. However, solutions that utilize more comprehensive environmental data, remote sensing information, and artificial intelligence (AI) algorithms may offer substantial advantages for rapid decision-making and extensive territory monitoring. The wide variety of spatial environmental parameters and the great diversity of geographical regions that influence wildfire occurrence complicate this task. Consequently, there is no unified approach for predicting wildfire occurrences using remote sensing data and AI techniques. The goal of this study is to explore the potential of predicting wildfire occurrences using various available environmental parameters - meteorological, geo-spatial, and anthropogenic - and machine learning (ML) algorithms. We developed a unified pipeline for data acquisition and subsequent ML-based algorithm development. The comprehensive analysis includes the following algorithms: Random Forest, XGBoost, Autoencoder, ConvLSTM, Attention Multilayer Perceptron, and RegNetX. In addition, we explore several metrics to assess the quality of developed models in case of highly imbalanced spatio-temporal data. To conduct the study, we collected a unique dataset covering several large regions in central Russia, incorporating more than 17,000 verified wildfire events over a period of 10 years. The findings underscore the necessity of developing individual ML models tailored to each region, taking into account the specific environmental features correlated with the probability of fire occurrence. The quality of the achieved models, as measured by F1-score, varies from 0.7 to 0.87 depending on the region, demonstrating the potential of integrating such algorithms into emergency response systems.

Amogel: a multi-omics classification framework using associative graph neural networks with prior knowledge for biomarker identification

The advent of high-throughput sequencing technologies, such as DNA microarray and DNA sequencing, has enabled effective analysis of cancer subtypes and targeted treatment. Furthermore, numerous studies have highlighted the capability of graph neural networks (GNN) to model complex biological systems and capture non-linear interactions in high-throughput data. GNN has proven to be useful in leveraging multiple types of omics data, including prior biological knowledge from various sources, such as transcriptomics, genomics, proteomics, and metabolomics, to improve cancer classification. However, current works do not fully utilize the non-linear learning potential of GNN and lack of the integration ability to analyse high-throughput multi-omics data simultaneously with prior biological knowledge. Nevertheless, relying on limited prior knowledge in generating gene graphs might lead to less accurate classification due to undiscovered significant gene-gene interactions, which may require expert intervention and can be time-consuming. Hence, this study proposes a graph classification model called associative multi-omics graph embedding learning (AMOGEL) to effectively integrate multi-omics datasets and prior knowledge through GNN coupled with association rule mining (ARM). AMOGEL employs an early fusion technique using ARM to mine intra-omics and inter-omics relationships, forming a multi-omics synthetic information graph before the model training. Moreover, AMOGEL introduces multi-dimensional edges, with multi-omics gene associations or edges as the main contributors and prior knowledge edges as auxiliary contributors. Additionally, it uses a gene ranking technique based on attention scores, considering the relationships between neighbouring genes. Several experiments were performed on BRCA and KIPAN cancer subtypes to demonstrate the integration of multi-omics datasets (miRNA, mRNA, and DNA methylation) with prior biological knowledge of protein-protein interactions, KEGG pathways and Gene Ontology. The experimental results showed that the AMOGEL outperformed the current state-of-the-art models in terms of classification accuracy, F1 score and AUC score. The findings of this study represent a crucial step forward in advancing the effective integration of multi-omics data and prior knowledge to improve cancer subtype classification.

Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training

Accurate real-time monitoring of underground conditions in coal mines is crucial for effective production management. However, limited computational resources and complex environmental conditions in mine shafts significantly impact the recognition and computational capabilities of detection models. This study utilizes a comprehensive dataset containing 117,887 images from five common underground mining tasks: mine personnel detection, large coal lump identification, conveyor chain monitoring, miner behavior recognition, and hydraulic support shield inspection. We propose the ESFENet backbone network, incorporating a Global Response Normalization (GRN) module to enhance feature capture stability while employing depthwise separable convolutions and HGRNBlock modules to reduce parameter volume and computational complexity. Building upon this foundation, we propose UCM-Net, a detection model based on the YOLO architecture. Furthermore, a self-supervised pre-training method is introduced to generate mine-specific pre-trained weights, providing the model with more semantic features. We propose utilizing the combined backbone and neck portions of the detection model as the encoder of an image-masking pre-training structure to strengthen feature acquisition and improve the performance of small models in self-supervised learning. Experimental results demonstrate that UCM-Net outperforms both baseline models and the state-of-the-art YOLOv12 model in terms of accuracy and parameter efficiency across the five mine datasets. The proposed architecture achieves 21.5% parameter reduction and 14.8% computational load decrease compared to baseline models while showing notable performance improvements of 1.3% (mAP50:95) and 0.8% (mAP50) in miner behavior recognition. The self-supervised pre-training framework effectively enhances training efficiency, enabling UCM-Net to attain an average mAP50 of 94.4% across all five datasets. The research outcomes can provide key technical support for coal mine safety monitoring and offer valuable technological insights for the public safety sector.

A Deep Learning Approach for Infant Pain Assessment Using Facial Expressions Through Convolutional Neural Network

This study presents a deep learning-based approach for assessing infant pain through facial expression analysis using Convolutional Neural Networks (CNNs). Given infants' inability to verbally articulate pain, reliable assessment methods are crucial in clinical nursing. To address this need, we developed a CNN model utilizing the COPE (Classification of Pain Expression) database. Our model achieved a test accuracy of 90.24%, with an average precision and recall of 87.58%, and an F1 score of 0.8758. Additionally, the model demonstrated high performance with an area under the curve of 0.9818 on the receiver operating characteristic curve. These results underscore the potential utility of CNNs for providing an objective pain assessment in clinical settings. However, the study acknowledges limitations, including a small sample size, the need for external validation, and ethical considerations. Future research should focus on expanding the dataset, conducting external validation, refining model architectures, and addressing ethical considerations to enhance performance and applicability. These efforts will advance infant pain management, ensure ethical integrity, and improve the overall quality of care.

MicrocrackAttentionNext: Advancing Microcrack Detection in Wave Field Analysis Using Deep Neural Networks Through Feature Visualization

Microcrack detection using deep neural networks (DNNs) through an automated pipeline using wave fields interacting with the damaged areas is highly sought after. However, these high-dimensional spatio-temporal crack data are limited. Moreover, these datasets have large dimensions in the temporal domain. The dataset presents a substantial class imbalance, with crack pixels constituting an average of only 5% of the total pixels per sample. This extreme class imbalance poses a challenge for deep learning models with different microscale cracks, as the network can be biased toward predicting the majority class, generally leading to poor detection accuracy for the under-represented class. This study proposes an asymmetric encoder-decoder network with an adaptive feature reuse block for microcrack detection. The impact of various activation and loss functions are examined through feature space visualisation using the manifold discovery and analysis (MDA) algorithm. The optimized architecture and training methodology achieved an accuracy of 87.74%.

Artificial intelligence accelerates the identification of nature-derived potent LOXL2 inhibitors

The role of LOXL2 in cancer has been widely demonstrated, but current therapies targeting LOXL2 are not yet fully developed. We believe that selective nature-derived inhibition of LOXL2 may provide a better therapeutic approach for the treatment of cancer. Therefore, we adopted a comprehensive approach combining deep learning and traditional computer-aided drug design methods to screen LOXL2 selective inhibitors. Bioactivity and affinity of the potential LOXL2 inhibitors were determined by molecular docking and virtual screening. At the same time, we experimentally tested the effect of potential LOXL2 inhibitors on cancer cells. Validation showed that it could inhibit proliferation and migration, promote apoptosis of CT26 cells, and reduce the expression level of LOXL2 protein. As a result, we identified a potent LOXL2 inhibitor: the natural product Forsythoside A, and demonstrated that Forsythoside A has an inhibitory effect on tumors.

Enzyme functional classification using artificial intelligence

Enzymes are essential for cellular metabolism, and elucidating their functions is critical for advancing biochemical research. However, experimental methods are often time consuming and resource intensive. To address this, significant efforts have been directed toward applying artificial intelligence (AI) to enzyme function prediction, enabling high-throughput and scalable approaches. In this review, we discuss advances in AI-driven enzyme functional annotation, transitioning from traditional machine learning (ML) methods to state-of-the-art deep learning approaches. We highlight how deep learning enables models to automatically extract features from raw data without manual intervention, leading to enhanced performance. Finally, we discuss the discovery of novel enzyme functions and generation of de novo enzymes through the integration of generative AIs and bio big data as future research directions.

Applications of AI in Predicting Drug Responses for Type 2 Diabetes

Unlabelled

Type 2 diabetes mellitus has seen a continuous rise in prevalence in recent years, and a similar trend has been observed in the increased availability of glucose-lowering drugs. There is a need to understand the variation in treatment response to these drugs to be able to predict people who will respond well or poorly to a drug. Electronic health records, clinical trials, and observational studies provide a huge amount of data to explore predictors of drug response. The use of artificial intelligence (AI), which includes machine learning and deep learning techniques, has the capacity to improve the prediction of treatment response in patients. AI can assist in the analysis of vast datasets to identify patterns and may provide valuable information on selecting an effective drug. Predicting an individual's response to a drug can aid in treatment selection, optimizing therapy, exploring new therapeutic options, and personalized medicine. This viewpoint highlights the growing evidence supporting the potential of AI-based methods to predict drug response with accuracy. Furthermore, the methods highlight a trend toward using ensemble methods as preferred models in drug response prediction studies.

A Hybrid Deep Learning and Improved SVM Framework for Real-Time Railroad Construction Personnel Detection with Multi-Scale Feature Optimization

Railroad construction sites are high-risk environments where monitoring personnel safety is critical for preventing accidents and enhancing construction efficiency. Traditional manual monitoring and image processing methods exhibit deficiencies in real-time performance and accuracy. This paper proposes a railway worker detection method based on improved support vector machines (ISVM), while using non-local mean noise reduction and histogram equalisation pre-processing techniques to optimise image quality to improve detection efficiency and accuracy. Multiscale features are then extracted with Inception v3 and combined with principal component analysis (PCA) for dimensionality reduction. Finally, an SVM classification algorithm is employed for personnel detection. To process small sample categories, data enhancement techniques (e.g., random flip and rotation) and K-fold cross-validation are applied to optimize the model parameters. The experimental results demonstrate that the ISVM method significantly improves accuracy and real-time performance compared to traditional detection methods and single deep learning models. This method provides technical support for railroad construction safety monitoring and effectively addresses personnel detection tasks in complex construction environments.

Global and regional patterns of soil metal(loid) mobility and associated risks

Soil contamination by metals and metalloids (metal[loid]s) is a global issue with significant risks to human health, ecosystems, and food security. Accurate risk assessment depends on understanding metal(loid) mobility, which dictates bioavailability and environmental impact. Here we show a theory-guided machine learning model that predicts soil metal(loid) fractionation across the globe. Our model identifies total metal(loid) content and soil organic carbon as primary drivers of metal(loid) mobility. We find that 37% of the world's land is at medium-to-high mobilization risk, with hotspots in Russia, Chile, Canada, and Namibia. Our analysis indicates that global efforts to enhance soil carbon sequestration may inadvertently increase metal(loid) mobility. Furthermore, in Europe, the divergence between spatial distributions of total and mobile metal(loid)s is uncovered. These findings offer crucial insights into global distributions and drivers of soil metal(loid) mobility, providing a robust tool for prioritizing metal(loid) mobility testing, raising awareness, and informing sustainable soil management practices.

Accelerating Multi-Objective Optimization of Composite Structures Using Multi-Fidelity Surrogate Models and Curriculum Learning

The optimization of multilayer composite structures requires balancing mechanical performance, economic efficiency, and computational feasibility. This study introduces an innovative approach that integrates Curriculum Learning (CL) with a multi-fidelity surrogate model to enhance computational efficiency in engineering design. A multi-fidelity strategy is introduced to generate training data efficiently, leveraging a high-fidelity finite element model for accurate simulations and a low-fidelity model to provide a larger dataset at reduced computational cost. Unlike conventional surrogate modeling approaches, the proposed method applies CL to iteratively refine the surrogate model, enabling step-by-step learning of complex structural patterns and improving prediction accuracy. Genetic algorithms (GAs) are then applied to optimize structural parameters while minimizing computational expense. The integration of CL and multi-fidelity modeling allows for a reduction in computational burden while preserving accuracy, demonstrating practical applicability in real-world structural design problems. The effectiveness of this methodology is validated by evaluating Pareto front quality using selected performance indicators. Results demonstrate that the proposed approach reduces optimization burden while achieving accurate predictions, highlighting the benefits of integrating surrogate modeling, multi-fidelity analysis, CL, and GAs for efficient composite structure optimization. This work contributes to the advancement of optimization methodologies by providing a scalable framework applicable to complex engineering problems requiring high computational efficiency.