Stress is a widespread concern that impacts human health with its silent progression, causing significant public health burdens and economic loss globally. Non-invasive wearable technology empowered by physiological signal monitoring can enable early warning systems for stress, alleviating some of the burdens, allowing on-time interventions, and thus significantly improving quality of life. This study used the heart rate and respiratory rate data from 34 participants. It evaluated the performance of the hybrid deep learning CNN-Transformer model and benchmarked it against deep learning convolutional neural networks (CNNs) and Transformer models, extreme gradient boosting (XGBoost) and random forest (RF) machine learning models, comprising a total of five AI models. To mitigate data imbalance and observe the efficacy of deep learning data augmentation techniques in physiological signals for stress monitoring, two generative adversarial network (GAN) models: conditional tabular GAN (CTGAN), copula GAN (CopGAN) and variational autoencoder (VAE) based model tabular VAE synthesiser (TVAES) had been employed. The modelling performance significantly improved when applying CTGAN and CopGAN, demonstrating the usefulness of synthetic data. The CNN-Transformer achieved an average accuracy of 77%, a precision of 87% and an AUC of 83%. The study applied leave-one- subject- out (LOSO CV) to prove the CNN-Transformer hybrid’s robustness for generalizability to perform well for unseen subjects. The study integrated explainable AI models, Shapley values (SHAP), and local interpretable model-agnostic explanations (LIME), as well as Monte Carlo Dropout for uncertainty quantification to bring confidence, trust and transparency to AI systems, taking a step closer to real-world deployment. Similar studies can also help in the detection of other disorders, such as anxiety and depression. • Comprehensive multi-domain feature extraction from heart rate and respiratory rate using wearable non-invasive sensor technology. • Comparative evaluation of three deep learning data augmentation techniques (CTGAN, TVAES and CopulaGAN) to mitigate data imbalance. • Comparative performance analysis of the hybrid DL model, CNN-Transformer model against benchmarking DL and ML models. • Implementing uncertainty quantification using Monte Carlo Dropout for confident and reliable model prediction assessment. • Application of xAI models LIME and SHAP for interpretability and feature importance.

Ensemble-based uncertainty quantification and decomposition of probabilistic surrogate models using Bayesian neural networks

Modeling study on NH3/CH4 co-combustion: sensitivity analysis, uncertainty quantification and mechanism optimization

Advancing fatigue crack growth prognosis in metallic structures: A physics-informed sequential attention approach with uncertainty quantification

Joint signal recovery and uncertainty quantification via the residual prior transform

Consensus computational immunogenicity modelling of botulinum neurotoxin serotypes: Cross-platform validation, uncertainty quantification, and relative risk assessment.

To study complex real-world phenomena using computer simulations, scientists often rely on ensemble datasets generated from multiple simulation runs with varying parameter configurations. This process can produce ensemble datasets with many members, making traditional data analysis pipelines impractical due to limited I/O bandwidth and disk capacity. Distribution-based data representations have been proposed as a promising solution. Processing data in situ to generate compact distribution-based representations not only alleviates the challenges of limited I/O bandwidth and disk capacity but also enables uncertainty quantification, thus mitigating the risk of misinterpretation. Nevertheless, distribution-based methods inherently sacrifice spatial information of data samples within the distribution, potentially reducing precision in the data analysis pipeline. To address this issue, we introduce a deep learning model to reconstruct data volume from the distribution representation. Instead of using a model that predicts a data block directly from its distribution representation, we propose a deep learning model based on the Sinkhorn operator and Gumbel trick that learns to map samples drawn from a distribution to spatial locations within the block. The deep learning model can support high-quality downstream data analysis and visualization, provide point-wise uncertainty quantification, and guarantee the distribution of the reconstructed data block follows the block's distribution representation.

Spatially correlated multi-state fragility via a warped Gaussian process

• Multi-fidelity delayed-acceptance MCMC for PDE-based Bayesian inverse problems • Offline-trained neural networks correct heterogeneous coarse solvers • No high-fidelity calls during sampling • Improved coarse-solver accuracy yields better mixing • Groundwater and reaction-diffusion benchmarks show computational savings Inverse uncertainty quantification (UQ) tasks such as parameter estimation are computationally demanding whenever dealing with physics-based models, and typically require repeated evaluations of complex numerical solvers. When partial differential equations are involved, full-order models such as those based on the Finite Element Method can make traditional sampling approaches like Markov Chain Monte Carlo (MCMC) computationally infeasible. Although data-driven surrogate models may help reduce evaluation costs, their utility is often limited by the expense of generating high-fidelity data. In contrast, low-fidelity data can be produced more efficiently, although relying on them alone may degrade the accuracy of the inverse UQ solution. To address these challenges, we propose a Multi-Fidelity Delayed Acceptance scheme for Bayesian inverse problems involving large-scale physics-based models. Extending the Multi-Level Delayed Acceptance framework, the method introduces multi-fidelity neural networks that combine the predictions of solvers of varying fidelity, with high-fidelity evaluations restricted to an offline training stage. During the online phase, likelihood evaluations are obtained by evaluating the coarse solvers and passing their outputs to the trained neural networks, thereby avoiding additional high-fidelity simulations. This construction allows heterogeneous coarse solvers to be incorporated consistently within the hierarchy, providing greater flexibility than standard Multi-Level Delayed Acceptance. The proposed approach improves the approximation accuracy of the low-fidelity solvers, leading to longer sub-chain lengths, better mixing, and accelerated posterior inference. The effectiveness of the strategy is demonstrated on two benchmark inverse problems involving (i) steady isotropic groundwater flow, (ii) an unsteady reaction–diffusion system, for which substantial computational savings are obtained.

Disentangled generative uncertainty-aware multi-modal diffusion segmentation of medical images.

Quantification and source identification of uncertainty in non-point source pollution loads

High-fidelity quantification of manufacturing-induced uncertainty in supersonic flow fields via deep autoencoder and spatially-adaptive polynomial chaos

Understanding root canal morphology is crucial for successful endodontic treatment; however, the anatomy of mandibular first premolars (M1Ps) remains one of the most variable and challenging aspects. The Vertucci classification provides a standardised framework for describing canal configurations; however, population-level data integrating multiple countries are scarce. This study aimed to evaluate the global distribution and determinants of Vertucci canal morphology in M1Ps using a Bayesian hierarchical model. Cone-beam computed tomography (CBCT) data of M1Ps from 21 countries were analysed. The Vertucci classification was used as the categorical outcome variable. The predictors included tooth side (34/44), voxel size, field of view (FOV), sex and age, with the country modelled as a random intercept. A Bayesian hierarchical multinomial logistic regression was fitted using the brms package (rstan backend) with weakly informative priors. Posterior estimates were expressed as odds ratios (OR) and 95% credible intervals (CrI), and model-based predicted probabilities were computed for each Vertucci type. Bayesian modelling estimated the posterior probability of Vertucci Type I configuration at 73.4% (95% CrI: 63.8%-81.5%). Non-Type I configurations showed lower but credible probabilities, including Type V (8.2%, 3.6%-15.9%), Type III (3.7%, 1.6%-7.7%), Type IV (2.9%, 1.2%-6.3%) and Type II (1.3%, 0.5%-3.1%). Unclassified canal patterns accounted for approximately one-tenth of the MnP1s (9.9%, 3.9%-19.2%). Substantial variability was observed between countries for non-Type I and unclassified configurations, whereas Type I remained consistently predominant. Sex and age exerted modest effects, whereas tooth side and field of view showed no meaningful associations. Increasing the voxel size was associated with a slight reduction in the probability of Type I and marginal increases in Type V and unclassified configurations. Although Vertucci Type I configuration predominates globally in MnP1s, clinically relevant non-Type I and unclassified canal patterns occur with non-negligible frequency and vary across populations. Bayesian hierarchical modelling enables the robust quantification of anatomical heterogeneity and uncertainty, supporting more reliable cross-country comparisons and cautious interpretation of less common canal configurations.

Integrating horizontal-vertical heterogeneity and environmental factors to unravel heavy metal distributions and phytoaccumulation in a soil-maize system.

Artificial intelligence for marine oil spill management: Recent advances and future directions.

Incorporating uncertainty quantification into deep-learning-based travel mode choice modeling: A Bayesian Neural Network approach and an uncertainty-guided active survey framework

Surrogate-based multilevel Monte Carlo methods for uncertainty quantification in the Grad-Shafranov free boundary problem

Diagnostic performance of AI-based EEG interpretation versus human clinical experts for epilepsy detection: systematic review and meta-analysis.

Pervasive sensing enables diverse wearable event detection (WED) applications, but deploying machine learning models on resource-constrained microcontrollers (MCUs) poses significant challenges, particularly in ensuring prediction reliability under data shifts or out-of-distribution (OOD) inputs. While Uncertainty quantification methods offer a way to assess this reliability, many are computationally prohibitive for MCUs, and detecting multiple events concurrently further exacerbates resource constraints. Addressing these combined challenges, this paper presents an uncertainty and resource-aware framework designed for reliable and efficient multi-event WED on MCUs, significantly extending our preliminary work. The proposed framework achieves this by integrating Evidential Deep Learning (EDL) for efficient, single-pass uncertainty estimation with a novel cascade learning architecture. This architecture promotes resource efficiency via: (i) intra-event sharing using uncertainty-aware early exits within a staged model (shallow, medium, deep), allowing simpler samples to terminate inference earlier; and (ii) inter-event sharing using a multi-head design where multiple event detectors share a common backbone, minimizing overhead. System efficiency is further enhanced through MCU-specific optimizations, including targeted architecture search, quantization, efficient uncertainty operator implementation using standard TensorFlow Lite Micro (TFLM) operations, and library footprint reduction. We conducted extensive experiments on four distinct wearable datasets (Oesense, KWS, ECG5000, and HHAR) and two MCU platforms (STM32F446ZE, STM32H747XI), comparing the proposed framework against strong baselines including Deep Ensembles and Vanilla EDL. Results demonstrate the proposed framework’s effectiveness, achieving competitive accuracy and uncertainty performance (e.g., up to 22% lower NLL than data augmentation) while drastically reducing resource consumption, offering up to 8.64 × faster inference, up to 8.57 × lower energy use, and 55% smaller memory footprint compared to ensemble methods. The proposed framework enables the deployment of reliable, uncertainty-aware multi-event detection on a wider range of low-power MCUs.

Accurate identification of mechanical properties is crucial for creating simplified equivalent models of complex aircraft structures but is often hindered by material heterogeneity, measurement errors, and multisource uncertainties. This study proposes an uncertainty-aware data-driven framework for full-field identification of spatially varying mechanical properties of aircraft structures, with integrated compensation for random, systematic, and gross measurement errors. Specifically, an enhanced U-Net model incorporating benchmark-calibrated and pixel-transformed datasets is designed to capture the nonlinear mapping from response fields to parameter distributions, enabling efficient online identification. Interval uncertainty quantification is achieved through statistical modeling of training residuals, providing reliable bounds that reflect uncertainties from data noise and model limitations. To enhance robustness under realistic measurement imperfections, random and gross errors are mitigated via noise-injected and data-missing-augmented training datasets, while systematic errors are corrected using a reduced-order basis approximation optimized through an active-learning-based surrogate modeling scheme. Numerical and experimental validations demonstrate that the proposed method accurately reconstructs full-field parameters, produces reliable uncertainty intervals, and maintains stability under various error conditions, highlighting its potential for structural equivalence modeling and online inverse characterization.