Predicting the pair correlation functions of silicate and borosilicate glasses using machine learning.

Machine learning (ML), including deep learning (DL), has become increasingly popular in the last few years due to its continually outstanding performance. In this context, we apply machine learning techniques to "learn" the microstructure using both supervised and unsupervised DL techniques. In particular, we focus (1) on the localization problem bridging (micro)structure (localized) property using supervised DL and (2) on the microstructure reconstruction problem in latent space using unsupervised DL. The goal of supervised and semi-supervised DL is to replace crystal plasticity finite element model (CPFEM) that maps from (micro)structure (localized) property, and implicitly the (micro)structure (homogenized) property relationships, while the goal of unsupervised DL is (1) to represent high-dimensional microstructure images in a non-linear low-dimensional manifold, and (2) to discover a way to interpolate microstructures via latent space associating with latent microstructure variables. At the heart of this report is the applications of several common DL architectures, including convolutional neural networks (CNN), autoencoder (AE), and generative adversarial network (GAN), to multiple microstructure datasets, and the quest of neural architecture search for optimal DL architectures.

Parkinson's disease (PD) presents a growing global health challenge, with early detection being crucial for effective management and treatment. This study seeks to develop an innovative machine learning (ML) framework for the early detection of PD by integrating advanced techniques for data preprocessing, dimensionality reduction, feature selection, and ensemble classification, aiming to significantly improve detection accuracy and timeliness. The research employs a robust ML pipeline, beginning with data preprocessing using mean imputation, standardization, min-max scaling, and SMOTE (Synthetic Minority Over-sampling Technique) to handle imbalanced data. Dimensionality reduction is achieved through Principal Component Analysis (PCA), while feature selection is performed using SelectKBest coupled with the ANOVA F-test to identify the most relevant features. Four ensemble methods—Random Forest, Gradient Boosting, XGBoost, and Support Vector Machine (SVM)—are evaluated for classification. Among the classifiers tested, the Gradient Boosting model stands out with an impressive accuracy of 0.9487, demonstrating its superior performance in PD detection. Integrating multiple preprocessing, dimensionality reduction, and feature selection techniques proves essential in optimizing model performance, highlighting the importance of a multifaceted approach in handling complex datasets. This research introduces a comprehensive ML framework that combines multiple advanced techniques in a streamlined process, significantly improving the early detection of Parkinson's disease. Ensemble methods, combined with strategic feature selection and data balancing techniques, offer a novel approach that could be applied to other neurodegenerative disorders, expanding its potential impact beyond PD detection.

In the last 10 years, various automated machine learning (AutoML) systems have been proposed to build end-to-end machine learning (ML) pipelines with minimal human interaction. Even though such automatically synthesized ML pipelines are able to achieve competitive performance, recent studies have shown that users do not trust models constructed by AutoML due to missing transparency of AutoML systems and missing explanations for the constructed ML pipelines. In a requirements analysis study with 36 domain experts, data scientists, and AutoML researchers from different professions with vastly different expertise in ML, we collect detailed informational needs for AutoML. We propose XAutoML , an interactive visual analytics tool for explaining arbitrary AutoML optimization procedures and ML pipelines constructed by AutoML. XAutoML combines interactive visualizations with established techniques from explainable artificial intelligence (XAI) to make the complete AutoML procedure transparent and explainable. By integrating XAutoML with JupyterLab , experienced users can extend the visual analytics with ad-hoc visualizations based on information extracted from XAutoML . We validate our approach in a user study with the same diverse user group from the requirements analysis. All participants were able to extract useful information from XAutoML , leading to a significantly increased understanding of ML pipelines produced by AutoML and the AutoML optimization itself.

A comparative study of machine learning methods for predicting the evolution of brain connectivity from a baseline timepoint

Chemical heuristics have been fundamental to the advancement of chemistry and materials science. These heuristics are typically established by scientists using knowledge and creativity to extract patterns from limited datasets. Machine learning offers opportunities to perfect this approach using computers and larger datasets. Here, we discuss the relationships between traditional heuristics and machine learning approaches. We show how traditional rules can be challenged by large-scale statistical assessment and how traditional concepts commonly used as features are feeding the machine learning techniques. We stress the waste involved in relearning chemical rules and the challenges in terms of data size requirements for purely data-driven approaches. Our view is that heuristic and machine learning approaches are at their best when they work together.

Emotion recognition plays a pivotal role in advancing Human-Computer Interaction (HCI) by enabling systems to understand and respond to human emotional states. Among various modalities, Speech Emotion Recognition (SER) stands out as a non-invasive, cost-effective, and temporally efficient approach for detecting emotions. This study leverages the RAVDESS dataset, which provides high-quality audio recordings for emotion classification. The proposed methodology involves preprocessing audio signals to remove noise, extracting temporal and spectral features in both time and frequency domains, and implementing machine learning models for multi-class emotion classification. The study evaluates and compares the performance of several classification models, including Random Forest (RF), Multilayer Perceptron (MLP), Support Vector Machine (SVM), Convolutional Neural Networks (CNN), and Decision Trees (DT). Experimental results demonstrate promising accuracy levels, highlighting the potential of these machine learning techniques in SER. This research contributes to the integration of SER within Brain-Computer Interfaces (BCI) and other emotion-aware applications, paving the way for enhanced interactive systems. Keywords – Speech Emotion Recognition -Machine Learning - Random Forest -Multilayer Perceptron - Support Vector Machine - Convolutional Neural Networks - Decision Tree - RAVDESS Dataset

Accurate prediction of significant wave height (SWH) is central to coastal ocean dynamics, wave–climate assessment, and operational marine forecasting, yet many high-performing machine-learning (ML) models remain opaque and weakly connected to underlying wave physics. We propose an explainable, feature engineering-guided ML framework for coastal SWH prediction that combines extremal wave statistics, temporal descriptors, and SHAP-based interpretation. Using 30 min buoy observations from a high-energy, wave-dominated coastal site off Australia’s Gold Coast, we benchmarked seven regression models (Linear Regression, Decision Tree, Random Forest, Gradient Boosting, Support Vector Regression, K-Nearest Neighbors, and Neural Networks) across four feature sets: (i) Base (Hmax, Tz, Tp, SST, peak direction), (ii) Base + Temporal (lags, rolling statistics, cyclical hour/month encodings), (iii) Base + a physics-informed Wave Height Ratio, WHR = Hmax/Hs, and (iv) Full (Base + Temporal + WHR). Model skill is evaluated for full-year, 1-month, and 10-day prediction windows. Performance was assessed using R2, RMSE, MAE, and bias metrics, with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) employed for multi-criteria ranking. Inclusion of WHR systematically improves performance, raising test R2 from a baseline range of ~0.85–0.95 to values exceeding 0.97 and reducing RMSE by up to 86%, with a Random Forest|Base + WHR configuration achieving the top TOPSIS score (1.000). SHAP analysis identifies WHR and lagged SWH as dominant predictors, linking model behavior to extremal sea states and short-term memory in the wave field. The proposed framework demonstrates how embedding simple, physically motivated features and explainable AI tools can transform black-box coastal wave predictors into transparent models suitable for geophysical fluid dynamics, coastal hazard assessment, and wave-energy applications.

Caveolae are plasma membrane invaginations whose formation requires caveolin-1 (Cav1), the adaptor protein polymerase I, and the transcript release factor (PTRF or CAVIN1). Caveolae have an important role in cell functioning, signaling, and disease. In the absence of CAVIN1/PTRF, Cav1 forms non-caveolar membrane domains called scaffolds. In this work, we train machine learning models to automatically distinguish between caveolae and scaffolds from single molecule localization microscopy (SMLM) data. We apply machine and deep learning algorithms to discriminate biological structures from SMLM data. Our work is the first that is leveraging machine and deep learning approaches to automatically identifying biological structures from SMLM data. In particular, we develop and compare three binary classification methods to identify whether or not a given 3D cluster of Cav1 proteins is a caveolae. The first uses a random forest classifier applied to 28 hand-picked features, the second uses a convolutional neural net (CNN) applied to a projection of the point clouds onto three planes, and the third uses a PointNet model, a recent development that can directly take point clouds as its input. We validate our methods on a dataset of super-resolution microscopy images of PC3 prostate cancer cells labeled for Cav1. Specifically, we have images from two cell populations: 10 PC3 and 10 CAVIN1/PTRF-transfected PC3 cells (PC3-PTRF cells) that form caveolae. We obtained a balanced set of 1714 different cellular structures. Our results show that both the random forest on hand selected features and the deep learning approach achieve high accuracy in distinguishing the intrinsic features of the caveolae and non-caveolae biological structures. More specifically, both random forest and deep CNN classifiers achieve classification accuracy reaching 94% on our test set, while the PointNet model only reached 83% accuracy. We also discuss the pros and cons of the different approaches. Additional information about individual files is in the accompanying CSV file (item_metadata.csv). This dataset was originally deposited in the Simon Fraser University institutional repository.

Hospital-acquired complications (HACs) have an adverse impact on patient recovery by impeding their path to full recovery and increasing healthcare costs. The aim of this study was to create a HAC risk prediction machine learning (ML) framework using hospital administrative data collections within North Metropolitan Health Service (NMHS), Western Australia. A retrospective cohort study was performed among 64,315 patients between July 2020 to June 2022 to develop an automated ML framework by inputting HAC and the healthcare site to obtain site-specific predictive algorithms for patients admitted to the hospital in NMHS. Univariate analysis was used for initial feature screening for 270 variables. Of these, 77 variables had significant relationship with any HAC. After excluding non-contemporaneous data, 37 variables were included in developing the ML framework based on logistic regression (LR), decision tree (DT) and random forest (RF) models to predict occurrence of four specific HACs: delirium, aspiration pneumonia, pneumonia and urinary tract infection. All models exhibited similar performance with area under the curve scores around 0.90 for both training and testing datasets. For sensitivity, DT and RF exceeded LR performance while on average, false positives were lowest for LR-based models. Patient's length of stay, Charlson Index, operation length and intensive care unit stay were common predictors. Integrating ML-based risk detection systems into clinical workflows can potentially enhance patient safety and optimise resource allocation. LR-based models exhibited best performance. We have successfully developed a "real-time" risk prediction model, where patient risk scores are calculated and reviewed daily.

The design and optimization of advanced semiconductor devices, such as multi gate MOSFETs, play a critical role in the miniaturization process of integrated circuits and the advancement of modern electronics. As device geometries shrink into the sub 10nm region, accurate modeling and calibration of transistor behavior become increasingly essential yet computationally challenging. This study presents a machine learning (ML) driven predictive modeling approach that uses TCAD extracted datasets to accurately predict the drain current (Id) of nanowire and nanoribbon FETs. Six ML models were trained and evaluated on TCAD data XGBoost, Random Forest, 1D Convolutional Neural Network (CNN), Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM), and TabNet. Among all models, the ensemble methods XGBoost and Random Forest performed the best, with R² scores of 0.9926 and 0.9549 and RMSE values of 2.1 × 10⁻⁴ A and 2.0 × 10⁻⁴ A, respectively. These results were significantly better than the deep learning models LSTM (R² = 0.6778) and 1D CNN (R² = 0.6997). These results demonstrate that tree-based models can generalize more effectively for structured TCAD data, challenging the prevailing assumption that deep learning models universally perform better in device level prediction tasks. The proposed ML framework significantly reduces the need for iterative TCAD simulations, enabling rapid and accurate prediction of device behavior. As a novel contribution, a fully integrated web-based application has been developed to deploy the best performing models for real time Id prediction and visualization.

Introduction: Heart failure (HF) in critical care settings exhibits complex interactions with glycemic dysregulation. Traditional metrics fail to capture dynamic glucose fluctuations linked to poor outcomes. This study defines ICU-specific glycemic targets and establishes a machine learning (ML) framework integrating novel glucose indicators to predict 28-day mortality. Methods: Cohort: 18,211 ICU HF patients (MIMIC-IV, 2008-2019). Analysis: 1. Glycemic Metrics: Calculated time-weighted average glucose (TWABG), time-in-range (TIR), time-below-range (TBR), time-above-range (TAR) 2. Association Studies: Restricted cubic splines (RCS) to identify optimal glycemic ranges. Multivariable Cox regression with adjustment and subgroup analysis with interaction testing. 3. ML Modeling: Dual-stage feature selection: Lasso regression and Boruta algorithm. Model development: Developed models using 14 algorithms; SHAP interpretation applied to best-performing model (highest ROC-AUC). Validation: SHAP analysis of an exemplar case demonstrated individualized mortality risk interpretation without formal model validation. Results: 1. Optimal Glycemic Ranges: Patients within the optimal TWABG range (85–133 mg/dL) demonstrated the lowest mortality (15.09%,P< 0.001), with improved survival associated with both higher time-in-range (TIR) proportion and absence of glycemic excursions (time-below-range [TBR] or time-above-range [TAR] exposure) 2. ML Prediction: Feature selection via Lasso regression identified 30 predictors. Boruta algorithm confirmed 29 significant features. The optimized XGBoost model predicted 28-day mortality with AUC 0.779 (0.762-0.796) Conclusions: Maintaining blood glucose within TWABG 85–133 mg/dL and maximizing TIR significantly reduces 28-day mortality in ICU HF patients, while avoiding excursions independently improves survival. Our ML framework, integrating dynamic glucose metrics with clinically selected features, accurately predicted mortality risk (AUC 0.779). Findings advocate ICU-specific glucose management and demonstrate ML’s potential for risk stratification and personalized interventions.

Breast cancer is one of the most prevalent and chronic disease that affect women. To overcome this disease, effective medical treatment is required. Early detection of the disease plays an important role for suitable medication and survival of patient. To identify the breast cancer in the patients, standard imaging technique mammography is used. Due to the subtle and varied nature of cancer tissues interpreting mammogram images can be a challenge to doctors. Machine learning (ML) and Deep Learning (DL) techniques offer promising solutions that provide efficient breast cancer detection from mammograms. In this review paper a comprehensive review of ML and DL algorithms and their applications in mammogram image analysis are presented. Various supervised and unsupervised learning techniques, such as convolutional neural networks (CNNs), support vector machines (SVMs), random forests, and other popular ML and DL models are discussed in paper. The integration of these DL methods that are efficiently used in image preprocessing techniques, feature extraction, and classification strategies. The overall survey focusses on various performance metrics, datasets, and benchmarks used in existing studies. Further the strengths and limitations of different approaches used by various researchers are identified. By understating current research trends this paper aims to contribute to the ongoing development of more accurate and reliable breast cancer detection systems using advanced ML techniques.

In recent years, a wide variety of automated machine learning (AutoML) methods have been proposed to generate end-to-end ML pipelines. While these techniques facilitate the creation of models, given their black-box nature, the complexity of the underlying algorithms, and the large number of pipelines they derive, they are difficult for developers to debug. It is also challenging for machine learning experts to select an AutoML system that is well suited for a given problem. In this paper, we present the Pipeline Profiler, an interactive visualization tool that allows the exploration and comparison of the solution space of machine learning (ML) pipelines produced by AutoML systems. PipelineProfiler is integrated with Jupyter Notebook and can be combined with common data science tools to enable a rich set of analyses of the ML pipelines, providing users a better understanding of the algorithms that generated them as well as insights into how they can be improved. We demonstrate the utility of our tool through use cases where PipelineProfiler is used to better understand and improve a real-world AutoML system. Furthermore, we validate our approach by presenting a detailed analysis of a think-aloud experiment with six data scientists who develop and evaluate AutoML tools.

Currently, there is no therapy targeting septic cardiomyopathy (SC), a key contributor to organ dysfunction in sepsis. In this study, we used a machine learning (ML) pipeline to explore transcriptomic, proteomic, and metabolomic data from patients with septic shock, and prospectively collected measurements of high-sensitive cardiac troponin and echocardiography. The purposes of the study were to suggest an exploratory methodology to identify and characterise the multiOMICs profile of (i) myocardial injury in patients with septic shock, and of (ii) cardiac dysfunction in patients with myocardial injury. The study included 27 adult patients admitted for septic shock. Peripheral blood samples for OMICS analysis and measurements of high-sensitive cardiac troponin T (hscTnT) were collected at two time points during the ICU stay. A ML-based study was designed and implemented to untangle the relations among the OMICS domains and the aforesaid biomarkers. The resulting ML pipeline consisted of two main experimental phases: recursive feature selection (FS) assessing the stability of biomarkers, and classification to characterise the multiOMICS profile of the target biomarkers. The application of a ML pipeline to circulate OMICS data in patients with septic shock has the potential to predict the risk of myocardial injury and the risk of cardiac dysfunction.

Epilepsy is a brain disease that causes people to have seizures over and over again. It is very hard to diagnose and treat. Finding epileptic seizures early is very important for getting help right away and improving the patient's result. This paper discussed about a way to find epileptic seizures using a mix of signal processing methods as feature extraction techniques, such as Principal Component Analysis (PCA) and Independent Component Analysis (ICA), along with different machine learning algorithms. The suggested framework is made up of several steps. First, noise and other errors are taken out of the raw EEG data. Next, PCA and ICA are used to pull out important features from the signals that have already been processed. These features are then fed into a number of machine learning methods to help them classify the data. The proposed framework uses Support Vector Machine (SVM), Random Forest, Convolutional Neural Networks (CNN), Long short-term memory (LSTM), CNN with BiLSTM, and Gradient Boosting, which are all machine learning methods used for correctness of proposed model. These methods are used to divide the EEG data into groups of seizures and non-seizures. A sample with EEG readings from people with epilepsy is used to test how well the system works. This dataset is used to train and test the framework. Performance measures like sensitivity, specificity, and accuracy are used to judge how well it works. According to the results, the suggested framework is very good at finding epileptic seizures. It combines PCA for feature extraction with a number of machine learning methods. The framework's flexible design makes it easy to add new features and methods and change how it works.&nbsp;