Search

Deep neural networks (DNNs) are prominent in predictive analytics for accurately forecasting target variables. However, inherent uncertainties necessitate constructing prediction intervals for reliability. The existing literature often lacks practical methodologies for creating predictive intervals, especially for time series with trends and seasonal patterns. This paper explicitly details a practical approach integrating dual-output Monte Carlo Dropout (MCDO) with DNNs to approximate predictive means and variances within a Bayesian framework, enabling forecast interval construction. The dual-output architecture employs a custom loss function, combining mean squared error with Softplus-derived predictive variance, ensuring non-negative variance values. Hyperparameter optimization is performed through a grid search exploring activation functions, dropout rates, epochs, and batch sizes. Empirical distributions of predictive means and variances from the MCDO demonstrate the results of the dual-output MCDO DNNs. The proposed method achieves a significant improvement in forecast accuracy, with an RMSE reduction of about 10% compared to the seasonal autoregressive integrated moving average model. Additionally, the method provides more reliable forecast intervals, as evidenced by a higher coverage proportion and narrower interval widths. A case study on Thailand's durian export data showcases the method's utility and applicability to other datasets with trends and/or seasonal components. [ABSTRACT FROM AUTHOR]

Autism Spectrum Disorder (ASD) is a chronic condition characterised by impairments in social interaction and communication. Early detection of ASD is desired, and there exists a demand for the development of diagnostic aids to facilitate this. A lightweight Involutional Neural Network (INN) architecture has been developed to diagnose ASD. The model follows a simpler architectural design and has less number of parameters than the state‐of‐the‐art (SOTA) image classification models, requiring lower computational resources. The proposed model is trained to detect ASD from eye‐tracking scanpath (SP), heatmap (HM), and fixation map (FM) images. Monte Carlo Dropout has been applied to the model to perform an uncertainty analysis and ensure the effectiveness of the output provided by the proposed INN model. The model has been trained and evaluated using two publicly accessible datasets. From the experiment, it is seen that the model has achieved 98.12% accuracy, 96.83% accuracy, and 97.61% accuracy on SP, FM, and HM, respectively, which outperforms the current SOTA image classification models and other existing works conducted on this topic. [ABSTRACT FROM AUTHOR]

Deep neural networks (DNNs) are prominent in predictive analytics for accurately forecasting target variables. However, inherent uncertainties necessitate constructing prediction intervals for reliability. The existing literature often lacks practical methodologies for creating predictive intervals, especially for time series with trends and seasonal patterns. This paper explicitly details a practical approach integrating dual-output Monte Carlo Dropout (MCDO) with DNNs to approximate predictive means and variances within a Bayesian framework, enabling forecast interval construction. The dual-output architecture employs a custom loss function, combining mean squared error with Softplus-derived predictive variance, ensuring non-negative variance values. Hyperparameter optimization is performed through a grid search exploring activation functions, dropout rates, epochs, and batch sizes. Empirical distributions of predictive means and variances from the MCDO demonstrate the results of the dual-output MCDO DNNs. The proposed method achieves a significant improvement in forecast accuracy, with an RMSE reduction of about 10% compared to the seasonal autoregressive integrated moving average model. Additionally, the method provides more reliable forecast intervals, as evidenced by a higher coverage proportion and narrower interval widths. A case study on Thailand’s durian export data showcases the method’s utility and applicability to other datasets with trends and/or seasonal components.

Abstract Autism Spectrum Disorder (ASD) is a chronic condition characterised by impairments in social interaction and communication. Early detection of ASD is desired, and there exists a demand for the development of diagnostic aids to facilitate this. A lightweight Involutional Neural Network (INN) architecture has been developed to diagnose ASD. The model follows a simpler architectural design and has less number of parameters than the state‐of‐the‐art (SOTA) image classification models, requiring lower computational resources. The proposed model is trained to detect ASD from eye‐tracking scanpath (SP), heatmap (HM), and fixation map (FM) images. Monte Carlo Dropout has been applied to the model to perform an uncertainty analysis and ensure the effectiveness of the output provided by the proposed INN model. The model has been trained and evaluated using two publicly accessible datasets. From the experiment, it is seen that the model has achieved 98.12% accuracy, 96.83% accuracy, and 97.61% accuracy on SP, FM, and HM, respectively, which outperforms the current SOTA image classification models and other existing works conducted on this topic.

Utility companies lose approximately 35 liters of water for every 100 produced due to incorrect, illegal connections and the poor condition of pipes. This study develops an intelligent model to detect leaks using the Kalman filter, BiLSTM neural networks, and the Monte Carlo Dropout algorithm. Using data from the Empresa de Acueductos y Alcantarillados de Bogotá (EAAB), Colombia, autocorrelation analysis, PCA, cluster analysis, ADF and Durbin-Watson tests, Hurst exponent, spectral analysis, and wavelet transform were performed. Then, Kalman filtering techniques were applied, and a BiLSTM architecture controlled with Monte Carlo dropout was implemented. The results showed an accuracy of 87.48% in training and 80.48% in validation. Temporal analysis revealed a stationary behavior in the flow series, and the decrease in spectral intensity around 0.25 Hz was related to pressure perturbations caused by leaks. A detailed evaluation of pressure and flow signals identified leak patterns with high precision, demonstrating the effectiveness of the wavelet spectrogram in detecting energy disturbances. The novelty of the study lies in the integration of advanced artificial intelligence and combinatorial optimization techniques to improve water resource management, allowing early and accurate detection of leaks, significantly improving compared to traditional methods. [ABSTRACT FROM AUTHOR]

Convolutional neural networks (CNNs) provide flexible function approximations for a wide variety of applications when the input variables are in the form of images or spatial data. Although CNNs often outperform traditional statistical models in prediction accuracy, statistical inference, such as estimating the effects of covariates and quantifying the prediction uncertainty, is not trivial due to the highly complicated model structure and overparameterization. To address this challenge, we propose a new Bayesian approach by embedding CNNs within the generalized linear models (GLMs) framework. We use extracted nodes from the last hidden layer of CNN with Monte Carlo (MC) dropout as informative covariates in GLM. This improves accuracy in prediction and regression coefficient inference, allowing for the interpretation of coefficients and uncertainty quantification. By fitting ensemble GLMs across multiple realizations from MC dropout, we can account for uncertainties in extracting the features. We apply our methods to biological and epidemiological problems, which have both high-dimensional correlated inputs and vector covariates. Specifically, we consider malaria incidence data, brain tumor image data, and fMRI data. By extracting information from correlated inputs, the proposed method can provide an interpretable Bayesian analysis. The algorithm can be broadly applicable to image regressions or correlated data analysis by enabling accurate Bayesian inference quickly. [ABSTRACT FROM AUTHOR]

This paper investigates uncertainty quantification (UQ) techniques in multi-class classification of chest X-ray images (COVID-19, Pneumonia, and Normal). We evaluate Bayesian Neural Networks (BNN) and the Deep Neural Network with UQ (DNN with UQ) techniques, including Monte Carlo dropout, Ensemble Bayesian Neural Network (EBNN), Ensemble Monte Carlo (EMC) dropout, across different evaluation metrics. Our analysis reveals that DNN with UQ, especially EBNN and EMC dropout, consistently outperform BNNs. For example, in Class 0 vs. All, EBNN achieved a UAcc of 92.6%, UAUC-ROC of 95.0%, and a Brier Score of 0.157, significantly surpassing BNN's performance. Similarly, EMC Dropout excelled in Class 1 vs. All with a UAcc of 83.5%, UAUC-ROC of 95.8%, and a Brier Score of 0.165. These advanced models demonstrated higher accuracy, better discriaminative capability, and more accurate probabilistic predictions. Our findings highlight the efficacy of DNN with UQ in enhancing model reliability and interpretability, making them highly suitable for critical healthcare applications like chest X-ray imageQ6 classification.

Plant phenology is the study of cyclical events in a plant life cycle such as leaf bud burst, flowering, and fruiting. In this article the problem of olive phenology prediction is addressed through the use of Deep Learning. Although Neural Networks have already been used in this area, to the best of our knowledge, this is the first implementation of Probabilistic Bayesian Neural Networks for olive phenology prediction. This architecture gives particular emphasis to estimating the model uncertainty, both aleatoric and epistemic. The Bayesian Inference method, more precisely the Variational Inference one, is compared with the Monte Carlo Dropout technique, which is known to be a less computationally intensive approximation of Variational Inference. For validation purposes, models performance is compared to the state-of-the-art results.

SignificanceOral cancer is one of the most prevalent cancers, especially in middle- and low-income countries such as India. Automatic segmentation of oral cancer images can improve the diagnostic workflow, which is a significant task in oral cancer image analysis. Despite the remarkable success of deep-learning networks in medical segmentation, they rarely provide uncertainty quantification for their output.AimWe aim to estimate uncertainty in a deep-learning approach to semantic segmentation of oral cancer images and to improve the accuracy and reliability of predictions.ApproachThis work introduced a UNet-based Bayesian deep-learning (BDL) model to segment potentially malignant and malignant lesion areas in the oral cavity. The model can quantify uncertainty in predictions. We also developed an efficient model that increased the inference speed, which is almost six times smaller and two times faster (inference speed) than the original UNet. The dataset in this study was collected using our customized screening platform and was annotated by oral oncology specialists.ResultsThe proposed approach achieved good segmentation performance as well as good uncertainty estimation performance. In the experiments, we observed an improvement in pixel accuracy and mean intersection over union by removing uncertain pixels. This result reflects that the model provided less accurate predictions in uncertain areas that may need more attention and further inspection. The experiments also showed that with some performance compromises, the efficient model reduced computation time and model size, which expands the potential for implementation on portable devices used in resource-limited settings.ConclusionsOur study demonstrates the UNet-based BDL model not only can perform potentially malignant and malignant oral lesion segmentation, but also can provide informative pixel-level uncertainty estimation. With this extra uncertainty information, the accuracy and reliability of the model’s prediction can be improved.

This study aims to enhance diagnostic capabilities for optimising the performance of the anaerobic sewage treatment lagoon at Melbourne Water's Western Treatment Plant (WTP) through a novel machine learning (ML)-based monitoring strategy. This strategy employs ML to make accurate probabilistic predictions of biogas performance by leveraging diverse real-life operational and inspection sensor and other measurement data for asset management, decision making, and structural health monitoring (SHM). The paper commences with data analysis and preprocessing of complex irregular datasets to facilitate efficient learning in an artificial neural network. Subsequently, a Bayesian mixture density neural network model incorporating an attention-based mechanism in bidirectional long short-term memory (BiLSTM) was developed. This probabilistic approach uses a distribution output layer based on the Gaussian mixture model and Monte Carlo (MC) dropout technique in estimating data and model uncertainties, respectively. Furthermore, systematic hyperparameter optimisation revealed that the optimised model achieved a negative log-likelihood (NLL) of 0.074, significantly outperforming other configurations. It achieved an accuracy approximately 9 times greater than the average model performance (NLL = 0.753) and 22 times greater than the worst performing model (NLL = 1.677). Key factors influencing the model's accuracy, such as the input window size and the number of hidden units in the BiLSTM layer, were identified, while the number of neurons in the fully connected layer was found to have no significant impact on accuracy. Moreover, model calibration using the expected calibration error was performed to correct the model's predictive uncertainty. The findings suggest that the inherent data significantly contribute to the overall uncertainty of the model, highlighting the need for more high-quality data to enhance learning. This study lays the groundwork for applying ML in transforming high-value assets into intelligent structures and has broader implications for ML in asset management, SHM applications, and renewable energy sectors. [ABSTRACT FROM AUTHOR]

Karst springs are essential drinking water resources, however, modeling them poses challenges due to complex subsurface flow processes. Deep learning models can capture complex relationships due to their ability to learn non‐linear patterns. This study evaluates the performance of the Transformer in forecasting spring discharges for up to 4 days. We compare it to the Long Short‐Term Memory (LSTM) Neural Network and a common baseline model on a well‐studied Austrian karst spring (LKAS2) with an extensive hourly database. We evaluated the models for two further karst springs with diverse discharge characteristics for comparing the performances based on four metrics. In the discharge‐based scenario, the Transformer performed significantly better than the LSTM for the spring with the longest response times (9% mean difference across metrics), while it performed poorer for the spring with the shortest response time (4% difference). Moreover, the Transformer better predicted the shape of the discharge during snowmelt. Both models performed well across all lead times and springs with 0.64–0.92 for the Nash–Sutcliffe efficiency and 10.8%–28.7% for the symmetric mean absolute percentage error for the LKAS2 spring. The temporal information, rainfall and electrical conductivity were the controlling input variables for the non‐discharge based scenario. The uncertainty analysis revealed that the prediction intervals are smallest in winter and autumn and highest during snowmelt. Our results thus suggest that the Transformer is a promising model to support the drinking water abstraction management, and can have advantages due to its attention mechanism particularly for longer response times. Key Points: The Transformer architecture was applied in karst hydrology for the first time, showing high performance for discharge forecastingMonte Carlo dropout revealed that the prediction intervals are smallest and cover the measured discharges best in winter and autumnThe high temporal resolution of the input data sets improved the forecasting performance [ABSTRACT FROM AUTHOR]

AbstractReactivity measurement methods, like the Amplified Source Method (ASM), link observable quantities to integral physics parameters characterizing subcritical assemblies (SCAs). These methods were mostly derived from point reactor kinetics, which assumes fundamental mode distribution. However, in SCAs, external sources cannot be neglected, leading to a nonideal response, such as the detector position dependence of measured $${k_{eff}}$$keff.This work investigates deterministic and probabilistic deep learning (DL) in determining $${k_{eff}}$$keff and kinetics/subcritical parameters using core map and foil/active detector responses as inputs, which distinguishes DL from neutronics codes. Convolutional neural networks surpassed dense neural networks with higher accuracy, while assigning a strong signature to appropriate core map features. Expansion into multi-input networks, which also process reaction rates, highlighted DL’s flexibility by accurate prediction regardless of reaction type.Uncertainty quantification of DL was done using Monte Carlo (MC) Dropout and Bayesian neural network (BNN). The results favored BNN over MC Dropout, showing greater improvement with increasing data. An assessment of ASM, applicable in a SCA at source equilibrium, showed a reactivity bias of up to −3.59%Δk/k (−4.86 $). In contrast, DL had a maximum bias of only 0.38%Δk/k (0.5 $). Underestimation by ASM represents a nonconservative scenario in criticality safety, while DL proved robust against spatial effects. This demonstrates DL’s potential in ensuring reactivity margins and a safe approach to criticality in reactor operation regimes where standard techniques can fail. [ABSTRACT FROM AUTHOR]

With the recent advancements in machine learning technology, the accuracy of autonomous driving object detection models has significantly improved. However, due to the complexity and variability of real-world traffic scenarios, such as extreme weather conditions, unconventional lighting, and unknown traffic participants, there is inherent uncertainty in autonomous driving object detection models, which may affect the planning and control in autonomous driving. Thus, the rapid and accurate quantification of this uncertainty is crucial. It contributes to a better understanding of the intentions of autonomous vehicles and strengthens trust in autonomous driving technology. This research pioneers in quantifying uncertainty in the YOLOv5 object detection model, thereby improving the accuracy and speed of probabilistic object detection, and addressing the real-time operational constraints of current models in autonomous driving contexts. Specifically, a novel probabilistic object detection model named M-YOLOv5 is proposed, which employs the MC-drop method to capture discrepancies between detection results and the real world. These discrepancies are then converted into Gaussian parameters for class scores and predicted bounding box coordinates to quantify uncertainty. Moreover, due to the limitations of the Mean Average Precision (MAP) evaluation metric, we introduce a new measure, Probability-based Detection Quality (PDQ), which is incorporated as a component of the loss function. This metric simultaneously assesses the quality of label uncertainty and positional uncertainty. Experiments demonstrate that compared to the original YOLOv5 algorithm, the M-YOLOv5 algorithm shows a 74.7% improvement in PDQ. When compared with the most advanced probabilistic object detection models targeting the MS COCO dataset, M-YOLOv5 achieves a 14% increase in MAP, a 17% increase in PDQ, and a 65% improvement in FPS. Furthermore, against the state-of-the-art probabilistic object detection models for the BDD100K dataset, M-YOLOv5 exhibits a 31.67% enhancement in MAP and a 125.6% increase in FPS.

Gas turbines are thermoelectric plants with various applications, such as large-scale electricity production, petrochemical industry, and steam generation. In order to optimize the operation of a gas turbine, it is necessary to develop system identification models that allow for the development of studies and analyses to increase the system's reliability. Current strategies for modeling complex and non-linear systems can be based on artificial intelligence techniques, using autoregressive neural networks of the NARX and LSTM type. In this context, this work aims to develop a model of a gas turbine capable of estimating the rotation speed of the turbine and simultaneously estimating the uncertainty associated with the estimation. These methodologies are based on artificial neural networks and the Monte Carlo dropout simulation method. The results were obtained from experimental data from a 215 MW gas turbine, getting the best model with a MAPE of 0.02% and an uncertainty associated with the turbine rotation speed of 2.2 RPM. [ABSTRACT FROM AUTHOR]

Recently, traffic flow prediction has gained popularity in the implementation of intelligent transportation systems. Most of the existing models for traffic flow prediction focus on increasing the prediction performance and providing fast predictions for real-time applications. In addition, they can reveal the integrity of a prediction when an actual value is provided. However, they cannot explain prediction uncertainty. Uncertainty has recently emerged as an important problem to be solved in deep learning. To address this issue, a Monte Carlo dropout method was proposed. This method estimates the uncertainty of a traffic prediction model. Using 5,729,640 traffic data points from Seoul, the model was designed to predict both the uncertainty and measurements. Notably, it performed better than the LSTM model. Experiments were conducted to show that the values predicted by the model and their uncertainty can be estimated together without significantly decreasing the performance of the model. In addition, a normality test was performed on the traffic flow uncertainty to confirm the normality, through which a benchmark for uncertainty was presented. Following these findings, the inclusion of uncertainty provides additional insights into our model, setting a new benchmark for traffic predictions, and enhancing the capabilities of intelligent transportation systems. [ABSTRACT FROM AUTHOR]

Building an orebody model is a key step in the design and operation of a mine because it provides the basis for follow-up mine decisions. Recently, it was shown that convolutional neural networks can successfully reproduce the manual geological interpretation of a complex ore deposit. The deep learning approach mitigates the shortcomings of a labor-intensive process that greatly limits the speed at which geological resources can be updated. However, convolutional neural network architectures lack the ability to measure the confidence of their predictions. In this study, we tried to assess the uncertainty of the boundaries of these domains so that the characterization of metal grades within them can account for this uncertainty. We explored and compared Monte Carlo Dropout and Bayesian neural networks to assess the uncertainty of deep convolutional neural network models trained to predict geological domains conditioned to drill-hole data. Monte Carlo Dropout uncertainty maps reflect the uncertainty in geological interpretations. The uncertainty is highest in areas where the interpreter/geologist had more difficulty delineating the boundaries of geological bodies. This is known as geological interpretation uncertainty. In contrast, Bayesian neural network uncertainty is visible depending on ore type frequency, complexity, and heterogeneity. Bayesian neural networks are able to better represent the uncertainty regarding the unknown. The application example here is a real case study of several ore types from a polymetallic sulfide orebody located in the south of Portugal. [ABSTRACT FROM AUTHOR]

Deep learning is widely used in many real-life applications. Despite their remarkable performance accuracies, deep learning networks are often poorly calibrated, which could be harmful in risk-sensitive scenarios. Uncertainty quantification offers a way to evaluate the reliability and trustworthiness of deep-learning-based model predictions. In this work, we introduced uncertainty quantification to our virtual research assistant recommender platform through both Monte Carlo dropout ensemble techniques. We also proposed a new formula to incorporate the uncertainty estimates into our recommendation models. The experiments were carried out on two different components of the recommender platform (i.e., a BERT-based grant recommender and a temporal graph network (TGN)-based collaborator recommender) using real-life datasets. The recommendation results were compared in terms of both recommender metrics (AUC, AP, etc.) and the calibration/reliability metric (ECE). With uncertainty quantification, we were able to better understand the behavior of our regular recommender outputs; while our BERT-based grant recommender tends to be overconfident with its outputs, our TGN-based collaborator recommender tends to be underconfident in producing matching probabilities. Initial case studies also showed that our proposed model with uncertainty quantification adjustment from ensemble gave the best-calibrated results together with the desirable recommender performance.

Training a deep learning model is always challenging as the data annotation requires expert knowledge, and is time consuming and laborious. To address this issue, the authors formulate an active learning framework to facilitate the training of deep learning models for performing concrete crack segmentation from images. The Monte Carlo dropout (MCDO) strategy, which requires no modification of deep learning models, is adopted to develop the uncertainty-based method to aid estimating the concrete crack features that the deep learning models are uncertain of, that is, feature representations that have not been well learned. Then, the informative data, that is, concrete crack images associated with high uncertainty score, are identified and retrieved for subsequent model training and optimization. The aforementioned processes can be repeated until all instances in the data pool are completely annotated or the target performance is attained. The feasibility of the proposed active learning framework is validated using an open-source concrete crack dataset. With only about 25% of training data, the deep learning model attains an intersection over union (IoU) of 0.930, which is about 99.2% of the score trained with all the training data (10,000 concrete crack images), demonstrating the capability of using sufficient amount of informative data to attain a promising result in concrete crack segmentation from images. [ABSTRACT FROM AUTHOR]

Data-driven algorithms have been widely applied in predicting tool wear because of the high prediction performance of the algorithms, availability of data sets, and advancements in computing capabilities in recent years. Although most algorithms are supposed to generate outcomes with high precision and accuracy, this is not always true in practice. Uncertainty exists in distinct phases of applying data-driven algorithms due to noises and randomness in data, the presence of redundant and irrelevant features, and model assumptions. Uncertainty due to noise and missing data is known as data uncertainty. On the other hand, model assumptions and imperfection are reasons for model uncertainty. In this paper, both types of uncertainty are considered in the tool wear prediction. Empirical mode decomposition is applied to reduce uncertainty from raw data. Additionally, the Monte Carlo dropout technique is used in training a neural network algorithm to incorporate model uncertainty. The unique feature of the proposed method is that it estimates tool wear as an interval, and the interval range represents the degree of uncertainty. Different performance measurement matrices are used to compare the proposed method. It is shown that the proposed approach can predict tool wear with higher accuracy. [ABSTRACT FROM AUTHOR]

When deep learning models are used to predict the probability distribution of future values, a task called probabilistic forecasting, they need to handle the epistemic and the aleatoric uncertainties. For the former, some are based on Monte Carlo dropout and have focused on prediction interval estimation assuming a normal distribution for the aleatoric uncertainty, without looking into general probabilistic forecasting tasks such as quantile regression or scenario generation. Time series in practice are usually not normally distributed, usually having a non-symmetrical distribution. So, these models need to be adapted with pre- and post-processing ad-hoc transformations to recover a normal distribution. We propose a supervised deep model based on Monte Carlo dropout that handles both sources of uncertainty and can be applied for multi-step probabilistic forecasting, including prediction interval estimation, quantile regression, and scenario generation. For the aleatoric uncertainty, our model allows changing the normality assumption easily to other families of distributions with the same number of parameters, such as Gamma, Weibull, and log-normal. For multi-step prediction, we use a procedure of re-injection of intermediate samples. To validate our proposal, we examine the behavior of the model on wind speed, wind power and electrical load forecasting, important tasks for the energy sector, as electrical grids present more uncertainty due to the penetration of variable renewable sources. We show that generating scenarios with our recurrent approach works better than directly estimating the distribution of all future values, also that considering the epistemic uncertainty makes the model more robust, and that changing the output distribution of the model is a property that may improve the metrics for specific datasets. [ABSTRACT FROM AUTHOR]

Recent performance breakthroughs in artificial intelligence (AI) and machine learning (ML), especially advances in deep learning, the availability of powerful and easy-to-use ML libraries (e.g., scikit-learn, TensorFlow, PyTorch), and increasing computational power, have led to unprecedented interest in AI/ML among nuclear engineers. For physics-based computational models, verification, validation, and uncertainty quantification (VVUQ) processes have been very widely investigated, and many methodologies have been developed. However, VVUQ of ML models has been relatively less studied, especially in nuclear engineering. This work focuses on uncertainty quantification (UQ) of ML models as a preliminary step of ML VVUQ, more specifically Deep Neural Networks (DNNs) because they are the most widely used supervised ML algorithm for both regression and classification tasks. This work aims at quantifying the prediction or approximation uncertainties of DNNs when they are used as surrogate models for expensive physical models. Three techniques for UQ of DNNs are compared, namely, Monte Carlo Dropout (MCD), Deep Ensembles (DE), and Bayesian Neural Networks (BNNs). Two nuclear engineering examples are used to benchmark these methods: (1) time-dependent fission gas release data using the Bison code and (2) void fraction simulation based on the Boiling Water Reactor Full-size Fine-Mesh Bundle Tests (BFBT) benchmark using the TRACE code. It is found that the three methods typically require different DNN architectures and hyperparameters to optimize their performance. The UQ results also depend on the amount of training data available and the nature of the data. Overall, all three methods can provide reasonable estimations of the approximation uncertainties. The uncertainties are generally smaller when the mean predictions are close to the test data while the BNN methods usually produce larger uncertainties than MCD and DE. [ABSTRACT FROM AUTHOR]

Although Deep Learning (DL) architectures have been used as efficient prediction tools in a variety of domains, they frequently do not care about the uncertainty in the predictions. This may prevent them from being used in practical applications. In seismic reservoir characterisation, predicting facies from seismic data is typically viewed as an inverse uncertainty quantification issue. The goal of the current study is to analyse the dependability of rock facies classification model in order to quantify the uncertainty while maintaining the high accuracy by using and evaluating monte carlo dropout based deep learning (MCDL), a computationally efficient technique. The proposed method is unique since it can quantify the epistemic uncertainty of the classified facies in blind or unseen well conditioned on Seismic attributes in the bayesian approximation achieved by MCDL framework. The findings show that MC dropout is successful in terms of accuracy and reliability, with a blind test F1-scores of 98% and 82% in predicting facies from synthetic and seismic datasets respectively. Moreover, the applications in a 2D section indicate that the internal regions of the seismic sections are generally classified with less epistemic uncertainty than their boundaries, as calculated from the different realizations of the MCDL network. For comparison, a plain DL and support vector machine (SVM) are also implemented and the findings suggest that our method outperformns the other models in comparison which indicates the potential of the model to be implemented in a robust rock facies classification.

Although Deep Learning (DL) models have been introduced in various fields as effective prediction tools, they often do not care about uncertainty. This can be a barrier to their adoption in real-world applications. The current paper aims to apply and evaluate Monte Carlo (MC) dropout, a computationally efficient approach, to investigate the reliability of several skip connection-based Convolutional Neural Network (CNN) models while keeping their high accuracy. To do so, a high-dimensional regression problem is considered in the context of subterranean fluid flow modeling using 376,250 generated samples. The results demonstrate the effectiveness of MC dropout in terms of reliability with a Standard Deviation (SD) of 0.012–0.174, and of accuracy with a coefficient of determination (R2) of 0.7881–0.9584 and Mean Squared Error (MSE) of 0.0113–0.0508, respectively. The findings of this study may contribute to the distribution of pressure in the development of oil/gas fields. [ABSTRACT FROM AUTHOR]

Unsupervised domain adaptation (UDA) aims to improve the prediction performance in the target domain under distribution shifts from the source domain. The key principle of UDA is to minimize the divergence between the source and the target domains. To follow this principle, many methods employ a domain discriminator to match the feature distributions. Some recent methods evaluate the discrepancy between two predictions on target samples to detect those that deviate from the source distribution. However, their performance is limited because they either match the marginal distributions or measure the divergence conservatively. In this paper, we present a novel UDA method that learns domain-invariant features that minimize the domain divergence. We propose model uncertainty as a measure of the domain divergence. Our UDA method based on model uncertainty (MUDA) adopts a Bayesian framework and provides an efficient way to evaluate model uncertainty by means of Monte Carlo dropout sampling. Experiment results on image recognition tasks show that our method is superior to existing state-of-the-art methods. We also extend MUDA to multi-source domain adaptation problems. [ABSTRACT FROM AUTHOR]

Gas turbines are thermoelectric plants with various applications, such as large-scale electricity production, petrochemical industry, and steam generation. In order to optimize the operation of a gas turbine, it is necessary to develop system identification models that allow for the development of studies and analyses to increase the system’s reliability. Current strategies for modeling complex and non-linear systems can be based on artificial intelligence techniques, using autoregressive neural networks of the NARX and LSTM type. In this context, this work aims to develop a model of a gas turbine capable of estimating the rotation speed of the turbine and simultaneously estimating the uncertainty associated with the estimation. These methodologies are based on artificial neural networks and the Monte Carlo dropout simulation method. The results were obtained from experimental data from a 215 MW gas turbine, getting the best model with a MAPE of 0.02% and an uncertainty associated with the turbine rotation speed of 2.2 RPM.

There is a growing interest of estimating the inherent uncertainty of photovoltaic (PV) power forecasts with probability forecasting methods to mitigate accompanying risks for system operators. This study aims to advance the field of probabilistic PV power forecast by introducing and extending deep Gaussian mixture density networks (MDNs). Using the sum of the weighted negative log likelihood of multiple Gaussian distributions as a minimizing objective, MDNs can estimate flexible uncertainty distributions with nearly all neural network structures. Thus, the advantages of advances in machine learning, in this case deep neural networks, can be exploited. To account for the epistemic (e.g., model) uncertainty as well, this study applies two ensemble approaches to MDNs. This is particularly relevant for industrial applications, as there is often no extensive (manual) adjustment of the forecast model structure for each site, and only a limited amount of training data are available during commissioning. The results of this study suggest that already seven days of training data are sufficient to generate significant improvements of 23.9% in forecasting quality measured by normalized continuous ranked probability score (NCRPS) compared to the reference case. Furthermore, the use of multiple Gaussian distributions and ensembles increases the forecast quality relatively by up to 20.5% and 19.5%, respectively. [ABSTRACT FROM AUTHOR]

The goal of this work is to develop accurate Machine Learning (ML) models for predicting the assembly axial neutron flux profiles in the SAFARI-1 research reactor, trained by measurement data from historical cycles. Affinity Propagation and k-means algorithms are used to identify clusters in the set of measured axial neutron flux profiles. Pair-counting-based and information-theoretic measures are applied to compare the clusterings. Deep Neural Network (DNN) and Gaussian Process (GP) ML models are then trained for different clusters, with prediction uncertainties quantified with the Monte Carlo Dropout. The proposed procedure improves the prediction accuracy for the control assemblies and reduces the prediction uncertainty, with axial flux shapes predicted by DNN and GP being very close and the overall accuracy becoming comparable to the fuel assemblies. • Machine Learning models for predicting the axial neutron flux profile are developed. • Models are trained with measured data from historical cycles of SAFARI-1 reactor. • Clustering algorithms are used to identify flux shape clusters in the dataset. • DNN and GP model are trained for each cluster and prediction uncertainty quantified. • The proposed procedure improves the prediction accuracy and reduces the uncertainty. [ABSTRACT FROM AUTHOR]

Plant phenology is the study of cyclical events in a plant life cycle such as leaf bud burst, flowering, and fruiting. In this article the problem of olive phenology prediction is addressed through the use of Deep Learning. Although Neural Networks have already been used in this area, to the best of our knowledge, this is the first implementation of Probabilistic Bayesian Neural Networks for olive phenology prediction. This architecture gives particular emphasis to estimating the model uncertainty, both aleatoric and epistemic. The Bayesian Inference method, more precisely the Variational Inference one, is compared with the Monte Carlo Dropout technique, which is known to be a less computationally intensive approximation of Variational Inference. For validation purposes, models performance is compared to the state-of-the-art results. • Novel approach to ML-based olive phenology modeling. • Model uncertainty estimation using Bayesian theory within a Deep Learning framework. • Comparison between Variational Inference and Monte Carlo Dropout implementations. • Proven usefulness of the target feature re-labeling during the training phase. • Competitive results with state-of-the-art models. [ABSTRACT FROM AUTHOR]

Data-driven algorithms have been widely applied in predicting tool wear because of the high prediction performance of the algorithms, availability of data sets, and advancements in computing capabilities in recent years. Although most algorithms are supposed to generate outcomes with high precision and accuracy, this is not always true in practice. Uncertainty exists in distinct phases of applying data-driven algorithms due to noises and randomness in data, the presence of redundant and irrelevant features, and model assumptions. Uncertainty due to noise and missing data is known as data uncertainty. On the other hand, model assumptions and imperfection are reasons for model uncertainty. In this paper, both types of uncertainty are considered in the tool wear prediction. Empirical mode decomposition is applied to reduce uncertainty from raw data. Additionally, the Monte Carlo dropout technique is used in training a neural network algorithm to incorporate model uncertainty. The unique feature of the proposed method is that it estimates tool wear as an interval, and the interval range represents the degree of uncertainty. Different performance measurement matrices are used to compare the proposed method. It is shown that the proposed approach can predict tool wear with higher accuracy.

Deep learning technologies, due to their advanced pattern extraction and recognition of high-dimensional data, have been widely adopted into multisensor-based fire detection systems. Since deep learning approaches can generate erroneous predictions due to incomplete training datasets, a retraining process over unseen observations is needed. However, storing a large amount of data from continuous multisensor streams and labeling them to create a retraining dataset are costly and time-consuming. In this paper, we propose an active learning framework based on an informative experience memory that is populated with meaningful retraining data by assessing the uncertainty of the data. In the proposed framework, the deep learning model predicts fire occurrence and estimates model uncertainty by taking advantage of a Bayesian neural network using Monte Carlo dropout. By storing only higher uncertain data points into the fixed-size informative experience memory and querying them to the system managers, the storage and labeling costs are minimized while improving performance. To evaluate our active learning framework with different neural network structures, we develop three Bayesian neural networks based on conventional classification networks, including the feedforward neural network, fully convolutional network, and long short-term memory. We further investigate various uncertainty assessment scoring methods for classification tasks such as entropy, BALD, variation ratios, and mean STD. Experiments on a real dataset show that the Bayesian FCN using the BALD assessment method has the highest performance gain with an F1 score of 0.95, with an improvement of 24% using only 700 data points.

We present a Model Uncertainty-aware Differentiable ARchiTecture Search ( $\mu $ DARTS) that optimizes neural networks to simultaneously achieve high accuracy and low uncertainty. We introduce concrete dropout within DARTS cells and include a Monte-Carlo regularizer within the training loss to optimize the concrete dropout probabilities. A predictive variance term is introduced in the validation loss to enable searching for architecture with minimal model uncertainty. The experiments on CIFAR10, CIFAR100, SVHN, and ImageNet verify the effectiveness of $\mu $ DARTS in improving accuracy and reducing uncertainty compared to existing DARTS methods. Moreover, the final architecture obtained from $\mu $ DARTS shows higher robustness to noise at the input image and model parameters compared to the architecture obtained from existing DARTS methods.

Past research on pedestrian trajectory forecasting mainly focused on deterministic predictions which provide only point estimates of future states. These future estimates can help an autonomous vehicle plan its trajectory and avoid collision. However, under dynamic traffic scenarios, planning based on deterministic predictions is not trustworthy. Rather, estimating the uncertainty associated with the predicted states with certain level of confidence can lead to robust path planning. Hence, the authors propose to quantify this uncertainty during forecasting using stochastic approximation which deterministic approaches fail to capture. The current method is simple and applies Bayesian approximation during inference to standard neural network architectures for estimating uncertainty. The authors compared the predictions between the probabilistic neural network (NN) models with the standard deterministic models. The results indicate that the mean predicted path of probabilistic models was closer to the ground truth when compared with the deterministic prediction. Further, the effect of stochastic dropout of weights and long-term prediction on future state uncertainty has been studied. It was found that the probabilistic models produced better performance metrics like average displacement error (ADE) and final displacement error (FDE). Finally, the study has been extended to multiple datasets providing a comprehensive comparison for each model.

The field of neuroimaging has increasingly sought to develop artificial intelligence-based models for neurological and neuropsychiatric disorder automated diagnosis and clinical decision support. However, if these models are to be implemented in a clinical setting, transparency will be vital. Two aspects of transparency are (1) confidence estimation and (2) explainability. Confidence estimation approaches indicate confidence in individual predictions. Explainability methods give insight into the importance of features to model predictions. In this study, we integrate confidence estimation and explainability approaches for the first time. We demonstrate their viability for schizophrenia diagnosis using resting state functional magnetic resonance imaging (rs-fMRI) dynamic functional network connectivity (dFNC) data. We compare two confidence estimation approaches: Monte Carlo dropout (MCD) and MC batch normalization (MCBN). We combine them with two gradient-based explainability approaches, saliency and layer-wise relevance propagation (LRP), and examine their effects upon explanations. We find that MCD often adversely affects model gradients, making it ill-suited for integration with gradient-based explainability methods. In contrast, MCBN does not affect model gradients. Additionally, we find many participant-level differences between regular explanations and the distributions of explanations for combined explainability and confidence estimation approaches. This suggests that a similar confidence estimation approach used in a clinical context with explanations only output for the regular model would likely not yield adequate explanations. We hope that our findings will provide a starting point for the integration of the two fields, provide useful guidance for future studies, and accelerate the development of transparent neuroimaging clinical decision support systems.

Lithium-ion batteries have been widely used for energy storage systems and vehicle industries, highly accurate remaining useful life (RUL) prediction of lithium-ion batteries is one of the key technologies on prognostics and health management. However, the uncertainty quantification and reliability of RUL prediction is ignored. To describe the uncertainty of the RUL prediction and avoid the over-fitting phenomenon, a model combining Monte Carlo Dropout (MC_dropout) and gated recurrent unit (GRU) is proposed. Firstly, the indirect health indicator is extracted and gray relation analysis (GRA) is used to analyze the relation with capacity Then, dropout method is combined with GRU model to avoid the phenomenon of gradient disappearance and over-fitting. The probability distribution of the prediction results and 95% confidence interval is obtained through the MC_dropout method, which established the uncertainty of the prediction model. Finally, propagation neural network based on particle swarm optimization (BPNN-PSO), the least-squares support vector machine (LS-SVM) is compared with the proposed method. The results show that the proposed model not only has superiority point prediction of RUL compared with other methods, but also can describe the uncertainty of RUL prediction.

The intermittence and fluctuation of renewable energy bring significant uncertainty to the power system, which enormously increases the operational risks of the power system. The development of efficient interval prediction models can provide data support for decision making and help improve the economy and reliability of energy interconnection operation. The performance of Bayesian deep learning models and Bayesian shallow neural networks in short-term interval prediction of photovoltaic power is compared in this study. Specifically, an LSTM Approximate Bayesian Neural Network model (ABNN-I) is built on the basis of the deep learning and Monte Carlo Dropout method. Meanwhile, a Feedforward Bayesian Neural Network (ABNN-II) model is introduced by Feedforward Neural Network and the Markov Chain Monte Carlo method. To better compare and verify the interval prediction capability of the ABNN models, a novel clustering method with three-dimensional features which include the number of peaks and valleys, the average power value, and the non-stationary measurement coefficient is proposed for generating sunny and non-sunny clustering sets, respectively. Results show that the ABNN-I model has an excellent performance in the field of photovoltaic short-term interval forecasting. At a 95% confidence level, the interval coverage from ABNN-I to ABNN-II can be increased by up to 3.1% and the average width of the interval can be reduced by 56%. Therefore, with the help of the high computational capacity of deep learning and the inherent ability to quantify uncertainty of the interval forecast from Bayesian methods, this research provides high-quality interval prediction results for photovoltaic power prediction and solves the problem of difficult modeling for over-fitting that exists in the training process, especially on the non-sunny clustering sets. [ABSTRACT FROM AUTHOR]

This paper introduces a theory-guided machine learning (TGML) framework, which combines a theoretical model (TM) and a machine learning (ML) algorithm to predict compaction density under cyclic loading. Several 1-D tests were conducted on uniformly graded fine sand compacted at varying moisture contents w , stress levels σ z and loading frequencies f , simulating the field compaction of materials using a vibratory roller. The laboratory compaction data were first analysed using a revised TM and an artificial neural network (ANN), and their performance was measured using mean absolute error (MAE). Next, the data were analysed using the TGML framework, which involves three different techniques. TGML1 increased the ML's ability to extrapolate (MAE improved from 2.2 × 10−3 to 1.2 × 10−3); TGML2 ensured ML and TM complemented each other to model observations better (MAE improved from 2.3 × 10−3 to 7.9 × 10−4); and TGML3 assisted in regularising the ML with an additional loss function which ensured the model followed the mechanistic understandings of the underlying physics (MAE improved from 9.2 × 10−3 to 2.7 × 10−3). Considering TGML3 during modelling is essential when dealing with noisy field datasets, and this is the highlight of this paper. TGML frameworks showed less error and lower model uncertainty, estimated using the novel Monte Carlo dropout technique. Furthermore, the developed TGML framework was used to demonstrate a termination criterion, i.e. the number of cycles of roller movement required to achieve the desired degree of compaction. Finally, an approach is proposed by which a simplified TM and ML model can estimate field compaction behaviour during roller movement. [ABSTRACT FROM AUTHOR]

This paper investigates uncertainty quantification (UQ) techniques in multi-class classification of chest X-ray images (COVID-19, Pneumonia, and Normal). We evaluate Bayesian Neural Networks (BNN) and the Deep Neural Network with UQ (DNN with UQ) techniques, including Monte Carlo dropout, Ensemble Bayesian Neural Network (EBNN), Ensemble Monte Carlo (EMC) dropout, across different evaluation metrics. Our analysis reveals that DNN with UQ, especially EBNN and EMC dropout, consistently outperform BNNs. For example, in Class 0 vs. All, EBNN achieved a U Acc of 92.6%, U AUC-ROC of 95.0%, and a Brier Score of 0.157, significantly surpassing BNN's performance. Similarly, EMC Dropout excelled in Class 1 vs. All with a U Acc of 83.5%, U AUC-ROC of 95.8%, and a Brier Score of 0.165. These advanced models demonstrated higher accuracy, better discriaminative capability, and more accurate probabilistic predictions. Our findings highlight the efficacy of DNN with UQ in enhancing model reliability and interpretability, making them highly suitable for critical healthcare applications like chest X-ray imageQ6 classification., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Whata, Dibeco, Madzima and Obagbuwa.)

Purpose: To develop a Bayesian convolutional neural network (BCNN) with Monte Carlo dropout sampling for metabolite quantification with simultaneous uncertainty estimation in deep learning–based proton MRS of the brain. Methods: Human brain spectra were simulated using basis spectra for 17 metabolites and macromolecules (N = 100 000) at 3.0 Tesla. In addition, actual in vivo spectra (N = 5) were modified by adjusting SNR and linewidth with increasing severity of spectral degradation (N = 50). A BCNN was trained on the simulated spectra to generate a noise‐free, line‐narrowed, macromolecule signal‐removed, metabolite‐only spectrum from a typical human brain spectrum. At inference, each input spectrum was Monte Carlo dropout sampled (50 times), and the resulting mean spectrum and variance spectrum were used for metabolite quantification and uncertainty estimation, respectively. Results: Using the simulated spectra, the mean absolute percent errors of the BCNN‐predicted metabolite content were < 10% for Cr, Glu, Gln, mI, NAA, and Tau (< 5% for Glu, NAA, and mI). For all metabolites, the correlations (r's) between the ground‐truth error and BCNN‐predicted uncertainty ranged 0.72–0.94 (0.83 ± 0.06; p < 0.001). Using the modified in vivo spectra, the extent of variation in the estimated metabolite content against the increasing severity of spectral degradation tended to be smaller with BCNN than with linear combination of model spectra (LCModel). Overall, the variation in metabolite content tended to be more highly correlated with the uncertainty from BCNN than with the Cramér‐Rao lower‐bounds from LCModel (0.938 ± 0.019 vs. 0.881 ± 0.057 [p = 0.115]). Conclusion: The BCNN with Monte Carlo dropout sampling may be used in deep learning–based MRS for the estimation of uncertainty in the machine‐predicted metabolite content, which is important in the clinical application of deep learning–based MRS. [ABSTRACT FROM AUTHOR]

Ensemble models have emerged as a powerful technique for improving robustness in medical image segmentation. However, traditional ensembles suffer from limitations such as under-confidence and over-reliance on poor performing models. In this work, we introduce an Adaptive Uncertainty-based Ensemble (AUE) model for tumor segmentation in histopathological slides. Our approach leverages uncertainty estimates from Monte Carlo dropout during testing to dynamically select the optimal pair of models for each whole slide image. The AUE model combines predictions from the two most reliable models (K-Net, ResNeSt, Segformer, Twins), identified through uncertainty quantification, to enhance segmentation performance. We validate the AUE model on the ACDC@LungHP challenge dataset, systematically comparing it against state-of-the-art approaches. Results demonstrate that our uncertainty-guided ensemble achieves a mean Dice score of 0.8653 and outperforms traditional ensemble techniques and top-ranked methods from the challenge by over 3 %. Our adaptive ensemble approach provides accurate and reliable lung tumor delineation in histopathology images by managing model uncertainty. • Adaptive uncertainty-based ensemble model (AUE) proposed for tumor segmentation. • AUE outperformed traditional ensemble models by a significant margin. • Utilizing uncertainty estimates enhances segmentation performance. [ABSTRACT FROM AUTHOR]

The proton exchange membrane fuel cell (PEMFC) is a promising power source, but the short lifespan and high maintenance cost restrict its development and widespread application. Performance degradation prediction is an effective technique to extend the lifespan and reduce the maintenance cost of PEMFC. This paper proposed a novel hybrid method for the performance degradation prediction of PEMFC. Firstly, considering the randomness of PEMFC degradation, a Wiener process model is established to describe the degradation of the aging factor. Secondly, the unscented Kalman filter algorithm is used to estimate the degradation state of the aging factor from monitoring voltage. Then, in order to predict the degradation state of PEMFC, the transformer structure is used to capture the data characteristics and fluctuations of the aging factor. To quantify the uncertainty of the predicted results, we also add the Monte Carlo dropout technology to the transformer to obtain the confidence interval of the predicted result. Finally, the effectiveness and superiority of the proposed method are verified on the experimental datasets.

Post‐hemorrhagic hydrocephalus (PHH) is a severe complication of intraventricular hemorrhage (IVH) in very preterm infants. PHH monitoring and treatment decisions rely heavily on manual and subjective two‐dimensional measurements of the ventricles. Automatic and reliable three‐dimensional (3D) measurements of the ventricles may provide a more accurate assessment of PHH, and lead to improved monitoring and treatment decisions. To accurately and efficiently obtain these 3D measurements, automatic segmentation of the ventricles can be explored. However, this segmentation is challenging due to the large ventricular anatomical shape variability in preterm infants diagnosed with PHH. This study aims to (a) propose a Bayesian U‐Net method using 3D spatial concrete dropout for automatic brain segmentation (with uncertainty assessment) of preterm infants with PHH; and (b) compare the Bayesian method to three reference methods: DenseNet, U‐Net, and ensemble learning using DenseNets and U‐Nets. A total of 41 T2‐weighted MRIs from 27 preterm infants were manually segmented into lateral ventricles, external CSF, white and cortical gray matter, brainstem, and cerebellum. These segmentations were used as ground truth for model evaluation. All methods were trained and evaluated using 4‐fold cross‐validation and segmentation endpoints, with additional uncertainty endpoints for the Bayesian method. In the lateral ventricles, segmentation endpoint values for the DenseNet, U‐Net, ensemble learning, and Bayesian U‐Net methods were mean Dice score = 0.814 ± 0.213, 0.944 ± 0.041, 0.942 ± 0.042, and 0.948 ± 0.034 respectively. Uncertainty endpoint values for the Bayesian U‐Net were mean recall = 0.953 ± 0.037, mean negative predictive value = 0.998 ± 0.005, mean accuracy = 0.906 ± 0.032, and mean AUC = 0.949 ± 0.031. To conclude, the Bayesian U‐Net showed the best segmentation results across all methods and provided accurate uncertainty maps. This method may be used in clinical practice for automatic brain segmentation of preterm infants with PHH, and lead to better PHH monitoring and more informed treatment decisions. [ABSTRACT FROM AUTHOR]

Purpose: Systematic prostate biopsy is widely used for cancer diagnosis. The procedure is blind to underlying prostate tissue micro-structure; hence, it can lead to a high rate of false negatives. Development of a machine-learning model that can reliably identify suspicious cancer regions is highly desirable. However, the models proposed to-date do not consider the uncertainty present in their output or the data to benefit clinical decision making for targeting biopsy. Methods: We propose a deep network for improved detection of prostate cancer in systematic biopsy considering both the label and model uncertainty. The architecture of our model is based on U-Net, trained with temporal enhanced ultrasound (TeUS) data. We estimate cancer detection uncertainty using test-time augmentation and test-time dropout. We then use uncertainty metrics to report the cancer probability for regions with high confidence to help the clinical decision making during the biopsy procedure. Results: Experiments for prostate cancer classification includes data from 183 prostate biopsy cores of 41 patients. We achieve an area under the curve, sensitivity, specificity and balanced accuracy of 0.79, 0.78, 0.71 and 0.75, respectively. Conclusion: Our key contribution is to automatically estimate model and label uncertainty towards enabling targeted ultrasound-guided prostate biopsy. We anticipate that such information about uncertainty can decrease the number of unnecessary biopsy with a higher rate of cancer yield. [ABSTRACT FROM AUTHOR]

Hate speech is an important problem in the management of user-generated content. To remove offensive content or ban misbehaving users, content moderators need reliable hate speech detectors. Recently, deep neural networks based on the transformer architecture, such as the (multilingual) BERT model, have achieved superior performance in many natural language classification tasks, including hate speech detection. So far, these methods have not been able to quantify their output in terms of reliability. We propose a Bayesian method using Monte Carlo dropout within the attention layers of the transformer models to provide well-calibrated reliability estimates. We evaluate and visualize the results of the proposed approach on hate speech detection problems in several languages. Additionally, we test whether affective dimensions can enhance the information extracted by the BERT model in hate speech classification. Our experiments show that Monte Carlo dropout provides a viable mechanism for reliability estimation in transformer networks. Used within the BERT model, it offers state-of-the-art classification performance and can detect less trusted predictions. [ABSTRACT FROM AUTHOR]

There is a growing interest of estimating the inherent uncertainty of photovoltaic (PV) power forecasts with probability forecasting methods to mitigate accompanying risks for system operators. This study aims to advance the field of probabilistic PV power forecast by introducing and extending deep Gaussian mixture density networks (MDNs). Using the sum of the weighted negative log likelihood of multiple Gaussian distributions as a minimizing objective, MDNs can estimate flexible uncertainty distributions with nearly all neural network structures. Thus, the advantages of advances in machine learning, in this case deep neural networks, can be exploited. To account for the epistemic (e.g., model) uncertainty as well, this study applies two ensemble approaches to MDNs. This is particularly relevant for industrial applications, as there is often no extensive (manual) adjustment of the forecast model structure for each site, and only a limited amount of training data are available during commissioning. The results of this study suggest that already seven days of training data are sufficient to generate significant improvements of 23.9% in forecasting quality measured by normalized continuous ranked probability score (NCRPS) compared to the reference case. Furthermore, the use of multiple Gaussian distributions and ensembles increases the forecast quality relatively by up to 20.5% and 19.5%, respectively.

This study proposes a practical solution for estimating the uncertainty in predicted values. One of the most crucial aspects of prediction is reserving one's judgment in the case of uncertainty and signaling an appropriate warning message to stakeholders. However, despite the increased popularity of machine learning models, measuring uncertainty in predicted values has not been the main concern, with land valuation being no exception to this practice. In this study, two very different machine learning models, a random forest and a Monte Carlo dropout neural network were used to estimate land prices. Gyeonggi province, South Korea, was chosen for the analysis, and variations in estimates from these models were exploited to measure predictive uncertainty. The uncertainty in land price was found to differ systematically depending on the land use category, with the largest magnitude of uncertainty for forest, followed by farming land, residential sites, and commercial sites. In general, the uncertainty in agricultural land price is higher than that in urban land price. Although property prices have been modeled extensively in the literature by a variety of methods, measurements of the uncertainty in predicted prices are limited. This study demonstrates that the uncertainty in land price can be described efficiently through machine learning models, and it is expected that the empirical findings identified in this study can be applied directly in valuation practices, promoting the rapid adoption of machine learning methods in the real estate industry. [ABSTRACT FROM AUTHOR]