Hasil untuk "machine learning"

Christophe Andrieu, Nando de Freitas, Arnaud Doucet et al.

Ming-Ying Lu, Jacky Chung-Hao Wu, Henry Horng-Shing Lu et al.

The global burden of hepatocellular carcinoma (HCC) has shifted from viral to nonviral etiologies. However, successful antiviral therapy does not fully eliminate the risk of HCC, underscoring the demand for more effective surveillance strategies. Current screening methods, such as semiannual ultrasonography and the measurement of α-fetoprotein levels, offer suboptimal sensitivity for early detection. A cost-effective, reliable surveillance approach remains an unmet need. The Barcelona Clinic Liver Cancer staging system provides a framework to guide HCC therapy; yet, some gray zone exists, particularly for patients with intermediate-stage disease. Although tyrosine kinase inhibitors and immunotherapies have transformed the therapeutic landscape, their efficacies vary among patients, highlighting the necessity for personalized treatment strategies. In response to these challenges, artificial intelligence (AI) approaches have emerged as transformative tools in healthcare. By processing complex, nonlinear relationships and uncovering hidden patterns in clinical data, AI methods offer capabilities beyond those of traditional statistical methods. Furthermore, AI-driven multi-omics analysis holds promise for identifying novel biomarkers, thereby advancing precision medicine for HCC patients. This review introduces the potential of AI applications in enhancing the diagnosis, treatment, and prognosis of HCC.

Ghenwa Chamouni, Filippo Lococo, Carolina Sassorossi et al.

IntroductionArtificial intelligence (AI) is increasingly integrating into the healthcare field, particularly in lung cancer care, including screening, diagnosis, treatment, and prognosis. While these applications offer promising advancements, they also raise complex challenges that must be addressed to ensure responsible implementation in clinical practice. This scoping review explores the ethical and legal aspects of AI applications in lung cancer.MethodsA search was conducted across PubMed, Scopus, Web of Science, Cochrane Library, PROSPERO, OAIster, and CABI. A total of 581 records were initially retrieved, of which 20 met the eligibility criteria and were included in the review. The PRISMA guidelines were followed.ResultsThe most frequently reported ethical concern was data privacy. Other recurrent issues included informed consent, no harm to patients, algorithmic bias and fairness, transparency, equity in AI access and use, and trust. The most frequently raised legal concerns were data protection and privacy, although issues relating to cybersecurity, liability, safety and effectiveness, the lack of appropriate regulation, and intellectual property law were also noted. Solutions proposed ranged from technical approaches to calls for regulatory and policy development. However, many studies lacked comprehensive legal analysis, and most included papers originated from high-income countries. This highlights the need for a broader global perspective.DiscussionThis review found that data privacy and protection are the most prominent ethical and legal concerns in AI applications for lung cancer care. Deep Learning (DL) applications, especially in diagnostic imaging, are closely tied to data privacy, lack of transparency, and algorithmic bias. Hybrid and multimodal AI systems raise additional concerns regarding informed consent and the lack of proper regulations. Ethical issues were more frequently addressed than legal ones, with limited consideration for global applicability, particularly in low- and lower middle-income countries. Although technical and policy solutions have been proposed, these remain largely unvalidated and fragmented, with limited real-world feasibility or scalability.

Xuening Lyu, Rimsa Goperma, Dandan Wang et al.

Abstract Background Niacin Skin-Flushing Response (NSR) has emerged as a promising objective biomarker for the precise diagnosis of mental disorders. However, its diagnostic potential has been constrained by the limitations of traditional statistical approaches. The advent of Artificial Intelligence (AI) offers a transformative opportunity to overcome these challenges. This study presents a novel contribution to the field by establishing an open-access dataset and developing advanced AI-driven tools to enhance the diagnostic accuracy of psychiatric disorders through NSR analysis. Methods This study introduces the world’s first open dataset specifically developed for AI studies of Niacin Skin-Flushing Response (NSR), a physiological biomarker associated with mental illnesses including depression, bipolar disorder, and schizophrenia. Leveraging this dataset, we developed an advanced Machine Learning (ML) approach designed for the broad diagnosis of mental disorders. Distinct from prior studies which are often limited to First Episode Schizophrenia and depend on specific devices, our approach champions device independence. The core of our methodology involves a novel algorithm featuring an Efficient-Unet based Deep Learning model for the precise segmentation of NSR areas. This segmentation is significantly enhanced by runtime data augmentation and trained on a robust train/validation/test dataset split. Subsequently, a Support Vector Machine (SVM) method is employed for psychiatric disorder classification utilizing feature vectors extracted from the segmentation of NSR areas with a 3-scale quantization. The SVM training incorporates 5-fold cross-validation, Synthetic Minority Over-sampling Technique (SMOTE) for managing class imbalance, and hyperparameter tuning to optimize balanced accuracy. Results The established dataset comprises 600 high-quality NSR images from 120 individuals, encompassing a diverse cohort of healthy controls and patients with various mental illnesses. The developed AI tools offer an objective, swift, and highly accurate approach that is demonstrably independent of the diagnosed condition or the specific device used for image acquisition. Comparative results demonstrate that the ML-based diagnostic approach achieves a sensitivity ranging from 60.0 to 65.0% and a specificity from 75.0 to 88.3% across various types of illnesses, further underscoring its broad applicability and device independence. Conclusions This research conclusively demonstrates the significant potential of advanced AI tools in achieving precise diagnosis of psychiatric disorders, potentially surpassing human capabilities in both speed and accuracy. With the provision of the proposed open dataset and the introduction of novel methodologies, this study marks substantial progress in developing an objective and accurate NSR-based screening process for a wide spectrum of psychiatric disorders. Its enhanced applicability and independence from specific devices hold profound potential to substantially advance mental health diagnostics and contribute to improved patient outcomes globally.

Nikolaos Pavlidis, Andreas Sendros, Theodoros Tsiolakis et al.

In an era of increasing reliance on digital health solutions, safeguarding user privacy has emerged as a paramount concern. Health applications often need to balance advanced AI functionalities with sufficient privacy measures to ensure user engagement. This paper presents the architecture of FLORA, a privacy-first ovulation-tracking application that leverages federated learning (FL), privacy-enhancing technologies (PETs), and blockchain to protect user data while delivering accurate and personalized health insights. Unlike conventional centralized systems, FLORA ensures that sensitive information remains on users’ devices, with predictive algorithms powered by local computations. Blockchain technology provides immutable consent tracking and model update transparency, further improving user trust. In addition, FLORA’s design incentivizes participation through a token-based reward system, fostering collaborative data contributions. This work illustrates how the integration of cutting-edge technologies creates a secure, scalable, and user-centric health application, setting a new standard for privacy-preserving digital health platforms.

Seyit Alperen Celtek, Seda Kul, A. Ozgur Polat et al.

The increasing adoption of electric vehicles (EVs) has led to the widespread implementation of battery swapping stations. However, ensuring fairness in battery pricing remains a significant challenge since variations in battery health and performance among swapped batteries can result in user dissatisfaction and operational inefficiencies. This paper introduces a novel approach to enhance fairness in battery swapping by integrating a machine learning-based real-time prediction model with a pricing strategy based on remaining useful life (RUL) estimation to address this issue. The proposed solution comprises a real-time RUL estimation system and a dynamic pricing mechanism that ensures fair pricing based on battery health and performance. This integrated approach aims to improve user satisfaction and the operational efficiency of swapping stations. The paper evaluates various machine learning algorithms for real-time RUL estimation regarding accuracy, computation time, and memory usage. The results suggest that XGBoost provides the most suitable balance between accuracy and efficiency, making it an effective solution for real-world applications. Comparative analysis shows that the XGBoost model outperforms the second-best method (Random Forest) with a lower error (3.50 vs 3.79) while maintaining competitive computational efficiency (9.75 vs 8.52 seconds) and memory usage (2.12 vs 2.32 MB) when solving a typical numerical case study problem. The proposed approach has the potential to accelerate the adoption of electric vehicles and contribute to sustainability goals by promoting efficient battery utilization and fair pricing mechanisms.

Samrat Chakraborty

Guillaume Mestrallet

Strategies for tackling cancer involve surgery, radiotherapy, chemotherapy, and immune checkpoint inhibitors (ICB). However, the effectiveness of ICB remains constrained, prompting the need for a proactive strategy to foresee treatment responses and resistances. This study undertook an analysis across diverse cancer patient cohorts (including melanoma, clear cell renal carcinoma, glioblastoma, bladder, and stomach cancers) subjected to various immune checkpoint blockade treatments. Surprisingly, our findings unveiled that over 38% of patients demonstrated resistance and persistent disease progression despite undergoing ICB intervention. To unravel the intricacies of resistance, we scrutinized the immune profiles of cancer patients experiencing ongoing disease progression and resistance post-ICB therapy. These profiles delineated multifaceted defects, including compromised macrophage, monocyte, and T cell responses, impaired antigen presentation, aberrant regulatory T cell (Tregs) responses, and an elevated expression of immunosuppressive and G protein-coupled receptor molecules (TGFB1, IL2RA, IL1B, EDNRB, ADORA2A, SELP, and CD276). Building upon these insights into resistance profiles, we harnessed machine learning algorithms to construct models predicting the response and resistance to ICB and developed the accompanying software. While previous work on glioblastoma with only one type of algorithm had an accuracy of 0.82, we managed to develop 20 models that provided estimates of future events of resistance or response in five cancer types, with accuracies ranging between 0.79 and 1, based on their distinct immune characteristics. In conclusion, our approach advocates for the personalized application of immunotherapy in cancer patients based on patient-specific attributes and computational models.

Chunyu Pan, Ying Ma, Lifei Wang et al.

Major Depressive Disorder (MDD) is a significant neurological condition associated with aberrations in brain functional networks. Traditional studies have predominantly analyzed these from a network topology perspective. However, given the brain’s dynamic and complex nature, exploring its mechanisms from a network control standpoint provides a fresh and insightful framework. This research investigates the integration of network controllability and machine learning to pinpoint essential biomarkers for MDD using functional magnetic resonance imaging (fMRI) data. By employing network controllability methods, we identify crucial brain regions that are instrumental in facilitating transitions between brain states. These regions demonstrate the brain’s ability to navigate various functional states, emphasizing the utility of network controllability metrics as potential biomarkers. Furthermore, these metrics elucidate the complex dynamics of MDD and support the development of precision medicine strategies that incorporate machine learning to improve the precision of diagnostics and the efficacy of treatments. This study underscores the value of merging machine learning with network neuroscience to craft personalized interventions that align with the unique pathological profiles of individuals, ultimately enhancing the management and treatment of MDD.

Mahbub Ul Islam Khan, Md. Ilius Hasan Pathan, Mohammad Mominur Rahman et al.

Electric vehicles (EVs) are commonly recognized as environmentally friendly modes of transportation. They function by converting electrical energy into mechanical energy using different types of motors, which aligns with the sustainable principles embraced by smart cities. The motors of EVs store and consume electrical power from renewable energy (RE) sources through interfacing connections using power electronics technology to provide mechanical power through rotation. The reliable operation of an EV mainly relies on the condition of interfacing connections in the EV, particularly the connection between the 3-<inline-formula> <tex-math notation="LaTeX">$\phi $ </tex-math></inline-formula> inverter output and the brushless DC (BLDC) motor. In this paper, machine learning (ML) tools are deployed for detecting and classifying the faults in the connecting lines from 3-<inline-formula> <tex-math notation="LaTeX">$\phi $ </tex-math></inline-formula> inverter output to the BLDC motor during operational mode in the EV platform, considering double-line and three-phase faults. Several machine learning-based fault identification and classification tools, namely the Decision Tree, Logistic Regression, Stochastic Gradient Descent, AdaBoost, XGBoost, K-Nearest Neighbour, and Voting Classifier, were tuned for identifying and categorizing faults to ensure robustness and reliability. The ML classifications were developed based on the datasets of healthy and faulty conditions considering the combination of six critical parameters that have significance in reliable EV operation, namely the current supplied to the BLDC motor from the inverter, the modulated DC voltage, output speed, and measured speed, as well as the output of the Hall-effect sensor. In addition, the superiority of the proposed fault detection and classification approaches using ML tools was assessed by comparing the detection and classification efficiency through some statistical performance parameter comparisons among the classifiers.

Silja Flenner, Adam Kubec, Christian David et al.

Nanotomography with hard X-rays is a widely used technique for high-resolution imaging, providing insights into the structure and composition of various materials. In recent years, tomographic approaches based on simultaneous illuminations of the same sample region from different angles by multiple beams have been developed at micrometre image resolution. Transferring these techniques to the nanoscale is challenging due to the loss in photon flux by focusing the X-ray beam. We present an approach for multi-beam nanotomography using a dual-beam Fresnel zone plate (dFZP) in a near-field holography setup. The dFZP generates two nano-focused beams that overlap in the sample plane, enabling the simultaneous acquisition of two projections from slightly different angles. This first proof-of-principle implementation of the dual-beam setup allows for the efficient removal of ring artifacts and noise using machine-learning approaches. The results open new possibilities for full-field multi-beam nanotomography and pave the way for future advancements in fast holotomography and artifact-reduction techniques.

Alireza Kasaie, Suchithra Rajendran

This study addresses patient unpunctuality, a major concern affecting patient waiting time, resource utilization, and quality of care. We develop and compare four machine learning models, including multinomial logistic regression, decision tree, random forest, and artificial neural network, to accurately predict patient arrival patterns and aid efficient scheduling. These models are analyzed using the explainable artificial intelligence approach and the Shapley additive explanations model, promoting comprehension and trust in our algorithmic results. Using three years of appointment data from a psychiatric clinic, we identify the travel distance, appointment lead time, patient’s age, Body Mass Index (BMI), and certain mental diagnoses as significant factors affecting the patient’s unpunctuality. Despite the good predictive potential of machine learning algorithms, no single model excels in all performance metrics. The study proposes implementing these machine learning techniques and the explainable artificial intelligence tool into the clinic’s appointment system as a decision support system to minimize patient unpunctuality.

Jing Liu, Richard G. J. Bellerby, Qing Zhu et al.

Abstract Sea surface salinity (SSS) is a master variable in oceanography and important to understand marine biogeochemical and physical processes. In the East China Sea (ECS), a random forest based regression ensemble model (RF) was developed to estimate the SSS with a spatial resolution of ∼1 km based on a large synchronous data set of in situ SSS observations, MODIS‐derived remote sensing reflectance (Rrs) and sea surface temperature (SST). The model showed the best performance when the Rrs(412), Rrs(488), Rrs(555), Rrs(667), SST and Julian day (JD) were used as inputs, with a root mean square error (RMSE) of 0.84, mean absolute error (MAE) of 0.31 and coefficient of determination (R2) of 0.81 for model training (N = 4,504), and a RMSE of 0.77, MAE of 0.30 and R2 of 0.86 for the model test (N = 1,153). The accuracy of the SSS model was examined using an independent data set during the period of 2020–2022 with a RMSE of 0.66 and MAE of 0.39 (N = 2,151). The interannual and seasonal signal of modeled SSS of the ECS, showed that important drivers of variability are the Changjiang discharge and the East‐Asian monsoon. Applications of the model to other Chinese marginal seas (Yellow and Bohai seas) showed good agreement in distribution patterns when compared with the estimated SSS from NASA Soil Moisture Active Passive. Once more empirical oceanographic data is made available, this robust model can be applied to other regions retraining the model with informed local data sets.

Ali Raza, Mohammad Rustom Al Nasar, Essam Said Hanandeh et al.

Kinematic motion detection aims to determine a person’s actions based on activity data. Human kinematic motion detection has many valuable applications in health care, such as health monitoring, preventing obesity, virtual reality, daily life monitoring, assisting workers during industry manufacturing, caring for the elderly. Computer vision-based activity recognition is challenging due to problems such as partial occlusion, background clutter, appearance, lighting, viewpoint, and changes in scale. Our research aims to detect human kinematic motions such as walking or running using smartphones’ sensor data within a high-performance framework. An existing dataset based on smartphones’ gyroscope and accelerometer sensor values is utilized for the experiments in our study. Sensor exploratory data analysis was conducted in order to identify valuable patterns and insights from sensor values. The six hyperparameters, tunned artificial indigence-based machine learning, and deep learning techniques were applied for comparison. Extensive experimentation showed that the ensemble learning-based novel ERD (ensemble random forest decision tree) method outperformed other state-of-the-art studies with high-performance accuracy scores. The proposed ERD method combines the random forest and decision tree models, which achieved a 99% classification accuracy score. The proposed method was successfully validated with the k-fold cross-validation approach.

Ning Zhang, Enxu Zhang, Fei Li

Soybean is a type of food crop with economic benefits. Whether they are damaged or not directly affects the survival and nutritional value of soybean plants. In machine learning, unbalanced data represent a major factor affecting machine learning efficiency, and unbalanced data refer to a category in which the number of samples in one category is much larger than that in the other, which biases the classification results towards a category with a large number of samples and thus affects the classification accuracy. Therefore, the effectiveness of the data-balancing method based on a convolutional neural network is investigated in this paper, and two balancing methods are used to expand the data set using the over-sampling method and using the loss function with assignable class weights. At the same time, to verify the effectiveness of the data-balancing method, four networks are introduced for control experiments. The experimental results show that the new loss function can effectively improve the classification accuracy and learning ability, and the classification accuracy of the DenseNet network can reach 98.48%, but the classification accuracy will be greatly reduced by using the data-augmentation method. With the binary classification method and the use of data-augmentation data sets, the excessive number of convolution layers will lead to a reduction in the classification accuracy and a small number of convolution layers can be used for classification purposes. It is verified that a neural network using a small convolution layer can improve the classification accuracy by 1.52% using the data-augmentation data-balancing method.

Mohammadhossein Karimizarchi, Davood Shishebori

Coronavirus disease 2019 or Covid-19, which is also called acute respiratory disease NCAV-2019 or commonly called corona, is a respiratory disease caused by acute respiratory syndrome coronavirus-2. Forecasting the number of new cases and deaths during todays can be a useful step in predicting the costs and facilities needed in the future. This study aims to model and predict new cases and deaths efficiently in the future. Nine popular forecasting techniques are tested on the data of Covid-19 in Yazd city as a case study. Using the evaluation criteria of mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), and the mean absolute percentage of error (MAPE) of the models are compared. According to the selected evaluation criteria, the results of the comprehensive analysis emphasize that the most efficient models are the ARIMA model for predicting the cumulative cases of hospitalization of Covid-19 and the Theta model for the cumulative cases of death. Also, the autoregressive neural network model has the worst performance among other models for both hospitalization and death cases.

Michael C. Robitaille, Jeff M. Byers, Joseph A. Christodoulides et al.

A self-supervised learning approach uses cellular motion between consecutive images to self-train a machine learning classifier for cell segmentation.