Quantifying uncertainty of mineral prediction using a novel Bayesian deep learning framework

Intelligent transportation management requires not only statistical information on users' mobility patterns, but also knowledge of their corresponding transportation modes. While GPS trajectories can be readily obtained from GPS sensors found in modern smartphones and vehicles, these massive geospatial data are neither automatically annotated nor segmented by transportation mode, subsequently complicating transportation mode identification. In addition, predictive uncertainty caused by the learned model parameters or variable noise in GPS sensor readings typically remains unaccounted for. To jointly address the above issues, we propose a Bayesian deep learning framework for unsupervised GPS trajectory segmentation. After unlabeled GPS trajectories are preprocessed into sequences of motion features, they are used in unsupervised training of a channel-calibrated temporal convolutional neural network for timestep-level transportation mode identification. At test time, we approximate variational inference via Monte Carlo dropout sampling, leveraging the mean and variance of the predicted distributions to classify each input timestep and estimate its predictive uncertainty, respectively. The proposed approach outperforms both its non-Bayesian variant and established GPS trajectory segmentation baselines on Microsoft's Geolife dataset without using any labels.

In this paper, we propose a Bayesian Deep Learning (BDL) framework to model uncertainty and predict the performance of terahertz (THz) biosensors with a graphene and molybdenum disulfide (MoS2) coating for AML biomarker detection. Although there have been studies on the individual advantage of these 2D materials for biosensing, a comparative analysis taking into account predictive uncertainty is still insufficient. To this end, we have generated a high-fidelity simulation dataset from full-wave EM simulations of DSSRR structures over the 0.1–2.5 THz frequency range. Realistic geometrical and dielectric modifications have been incorporated to mimic bio-sensing conditions. An approach based on a Bayesian Neural Network (BNN) with Monte Carlo dropout was employed for predicting sensitivity, Q-factor, resonance shift, and absorption, along with the estimation of aleatoric, as well as epistemic, uncertainty. Our results demonstrated a trade-off between material types: MoS2 sensors showed higher sensitivity (3548 GHz/RIU) but with a larger prediction uncertainty range of ±118 GHz/RIU; on the other hand, graphene-based sensors exhibited a better spectral resolution (Q = 48.5) and a more reliable QV prediction range of ±42 GHz/RIU. The uncertainty study further revealed that graphene demonstrated a predominance for aleatoric uncertainty (68%), classifying them as predictable physical characteristics, while MoS2 presents a higher epistemic one (55%), indicating sensitivity towards underrepresented design cases. We present a material selection algorithm based on utility that balances sensitivity, resolution, and uncertainty, demonstrating that MoS2 is the best choice for early screening, while graphene is more suitable for high-precision diagnostics. This study offers a scalable and reliable AI framework for quick, uncertainty-aware optimization of THz biosensors, which is directly applicable to clinical diagnostics and 2D-material-based photonic design.

This study, for the first time, presented a novel Bayesian inference deep learning model based on a dilated convolutional ResNet framework, utilizing acoustic emission (AE) signals as inputs to quantify the uncertainty of the compression after impact (CAI) strength in composites. The output of the deep learning model included the mean CAI and the predictive aleatoric uncertainty. The epistemic uncertainty of the model was estimated by randomly sampling the input multiple times. AE signals were divided into a complete training dataset and a partially missing training dataset to investigate the impact of data insufficiency on the uncertainty of deep learning models. The results demonstrate that the deep learning model can provide highly accurate predictions of CAI performance. It is worth noting that on incomplete datasets, the epistemic uncertainty of the predictions increased significantly. To evaluate the effect of noise on the prediction of CAI strength, Gaussian noise with signal‐to‐noise ratios (SNRs) of 10, 5, and 2 dB was introduced during the prediction phase. The Bayesian posterior probability density function becomes lower and wider under the same impact energy as external noise increases. The model exhibited excellent regression performance under SNRs of 10 and 5 dB, validating the robustness of the proposed framework. As noise levels increased, the range of aleatoric uncertainty in the predictions expanded.Highlights A dilated convolution Bayesian deep learning framework for uncertainty quantification of compression after impact strength in laminated composites is proposed. The prediction and uncertainty quantification performance of the framework is evaluated. The effect of data missing and background noise on the uncertainty is investigated.

A global sensitivity analysis and Bayesian inference framework was developed for improving the parameterization and predictability of a monthly time step process‐based biogeochemistry model. Using a Latin Hypercube sampler and an existing Terrestrial Ecosystem Model (TEM), a set of 500,000 Monte Carlo ensemble simulations was conducted for a black spruce forest ecosystem. A global sensitivity analysis was then conducted to identify the key model parameters and examine the interaction structures among TEM parameters. Bayesian inference analysis was also performed using these ensemble simulations and eddy flux data of carbon, latent heat flux, and MODIS gross primary production (GPP) to reduce the uncertainty of parameter estimation and prediction of TEM. We found that (1) the simulated carbon fluxes are mostly affected by parameters of the maximum rate of photosynthesis (CMAX), the half‐saturation constant for CO2 uptake by plants (kc), the half‐saturation constant for Photosynthetically Active Radiation used by plants (ki), and the change in autotrophic respiration due to 10°C temperature increase (RHQ10); (2) the effect of parameters on seasonal carbon dynamics varies from one parameter to another during a year; (3) to well constrain the uncertainties of TEM predictions and parameters using the Bayesian inference technique, at least two different fluxes of NEP, GPP, and ecosystem respiration (RESP) are required; and (4) different assumptions of the error structures of the flux data used in the Bayesian inference analysis result in different uncertainty bounds of the posterior parameters and model predictions. We further found that, using the Bayesian framework and eddy flux and satellite data, the uncertainty of simulated carbon fluxes has been remarkably reduced. The developed global sensitivity analysis and Bayesian framework could further be used to analyze and improve the predictability and parameterization of relatively coarse time step biogeochemistry models when the eddy flux and satellite data are available for other terrestrial ecosystems.

In this paper we present a probabilistic formalization of the instance-based learning approach. In our Bayesian framework, moving from the construction of an explicit hypothesis to a data-driven instancebased learning approach, is equivalent to averaging over all the (possibly infinitely many) individual models. The general Bayesian instance-based learning framework described in this paper can be applied with any set of assumptions defining a parametric model family, and to any discrete prediction task where the number of simultaneously predicted attributes is small, which includes for example all classification tasks prevalent in the machine learning literature. To illustrate the use of the suggested general framework in practice, we show how the approach can be implemented in the special case with the strong independence assumptions underlying the so called Naive Bayes classifier. The resulting Bayesian instance-based classifier is validated empirically with public domain data sets and the results are compared to the performance of the traditional Naive Bayes classifier. The results suggest that the Bayesian instancebased learning approach yields better results than the traditional Naive Bayes classifier, especially in cases where the amount of the training data is small.

3D-grid-based chemical transport models, such as the Community Multiscale Air Quality (CMAQ) modeling system, have been widely used for predicting concentrations of ambient air pollutants. However, typical horizontal resolutions of nationwide CMAQ simulations (12 × 12 km2) cannot capture local-scale gradients for accurately assessing human exposures and environmental justice disparities. In this study, a Bayesian ensemble machine learning (BEML) framework, which integrates 13 learning algorithms, was developed for downscaling CMAQ estimates of ozone daily maximum 8 h averages to the census tract level, across the contiguous US, and was demonstrated for 2011. Three-stage hyperparameter tuning and targeted validations were designed to ensure the ensemble model's ability to interpolate, extrapolate, and capture concentration peaks. The Shapley value metric from coalitional game theory was applied to interpret the drivers of subgrid gradients. The flexibility (transferability) of the 2011-trained BEML model was further tested by evaluating its ability to estimate fine-scale concentrations for other years (2012-2017) without retraining. To demonstrate the feasibility of using the BEML approach to strictly "data-limited" situations, the model was applied to downscale CMAQ outputs for a future-year scenario-based simulation that considers effects of variations in meteorology associated with climate change.

Refined sparse Bayesian learning configuration for stochastic response analysis

Bayesian deep learning applied to diabetic retinopathy with uncertainty quantification.

In Orthogonal Frequency Division Multiplexing (OFDM) systems, the technique used to estimate and track the time-varying multipath channel is critical to ensure reliable, high data rate communications. It is recognized that wireless channels often exhibit a sparse structure, especially for wideband and ultra-wideband systems. In order to exploit this sparse structure to reduce the number of pilot tones and increase the channel estimation quality, the application of compressed sensing to channel estimation is proposed. In this article, to make the compressed channel estimation more feasible for practical applications, it is investigated from a perspective of Bayesian learning. Under the Bayesian learning framework, the large-scale compressed sensing problem, as well as large time delay for the estimation of the doubly selective channel over multiple consecutive OFDM symbols, can be avoided. Simulation studies show a significant improvement in channel estimation MSE and less computing time compared to the conventional compressed channel estimation techniques.

Process monitoring plays an important role in ensuring the safety and stable operation of equipment in a large-scale process. This paper proposes a novel data-driven process monitoring framework, termed the ensemble adaptive sparse Bayesian transfer learning machine (EAdspB-TLM), for nonlinear fault diagnosis. The proposed framework has the following advantages: Firstly, the probabilistic relevance vector machine (PrRVM) under Bayesian framework is re-derived so that it can be used to forecast the plant operating conditions. Secondly, we extend the PrRVM method and assimilate transfer learning into the sparse Bayesian learning framework to provide it with the transferring ability. Thirdly, the source domain (SD) data are re-enabled to alleviate the issue of insufficient training data. Finally, the proposed EAdspB-TLM framework was effectively applied to monitor a real wastewater treatment process (WWTP) and a Tennessee Eastman chemical process (TECP). The results further demonstrate that the proposed method is feasible.

Background: Immune checkpoint inhibitors (ICIs) targeting PD-1 or PD-L1 have dramatically improved survival of patients with gastric cancer (GC). However, recent clinical trials, including KEYNOTE-059, KEYNOTE-061, and ATTRACTION-2, show that the objective response rate (ORR) for ICIs is still less than 20% in patients with advanced GC. Therefore, predicting who will respond to ICIs in advance is essential to make better treatment plans for GC patients. There are potential biomarkers to predict ICI response, but most of them can only be applied to a limited population of patients. For example, Kim ST., et al., (PMID: 30013197) show that patients with MSI-high or EBV-positive responded well to ICIs, but these subtypes account for less than 30 percent of GC patients (PMID:25079317). In this study, we propose a prediction model of ICI response using the 32 GC-specific genes that we identified from mutation data (PMID: 26635139) and showed their close association with survival and ICI response in GC (PMID: 35140202; PMID: 36508166). In addition, since the number of available GC-ICI samples is small, we design our prediction model in Bayesian and ensemble learning frameworks which are robust to small sample size problems. Methods: We collected mRNA sequencing data from GC-ICI treated patients from our institutions and publications (PMID: 30013197; PMID: 35254400; PMID: 33846173). We used 108 samples previously included in our previous work (PMID: 36508166) as the training data and 27 samples (13 responders and 14 non-responders) from (PMID: 3384617) as the test data. We built and trained a classifier to separate responders from non-responders using the expressions of the 32 genes. We proposed to use a Deep Gaussian process (GP) model, a multi-layered extension of a GP that is itself a Bayesian method. We optimized the parameters of each GP model in variational inference. To implement ensemble learning, we trained multiple DGP models with different initializations of the model parameters and data splitting (training/validation sets). For each training run, the best model was selected based on the prediction performance on the validation set, measured by AUC. The final prediction for a test sample was made by combining the prediction results of all the DGP models by taking the average. Results: We implemented the DGP-based prediction model using PyTorch in Python, and the codes will be publicly available on the GitHub repository (https://github.com/hwanglab/GC_ICI_response_prediction). We achieved a prediction performance of 0.813 AUC on the test data (sensitivity = 0.769 and specificity = 0.714). Conclusion: We propose the prediction model of ICI response, which can handle small sample size problems in the framework of Bayesian and ensemble learning. The experimental results using the GC-ICI data demonstrate that the 32 genes can be used as predictive markers for ICI response in GC. Citation Format: Sunho Park, Minji Kim, Changjin Hong, Tae Hyun Hwang. A deep Gaussian process (DGP)-based prediction model for immunecheckpoint inhibitor (ICI) response in gastric cancer (GC) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2316.

Chronic knee conditions and musculoskeletal injuries are prevalent and pose significant challenges to global healthcare systems. Current diagnostic methods, such as MRI and X-ray, provide valuable insights but are limited by high costs and the inability to continuously monitor knee health. In response to these challenges, the proposed work makes the following contributions. Firstly, a novel wearable knee health monitoring system is developed, incorporating a flexible composite piezoelectric film that captures subtle patellar displacements as early indicators of knee joint abnormalities. Secondly, a Bayesian machine learning framework incorporating both expert knowledge and diagnostic data is proposed to address the potential limitations of current machine learning-based diagnostic methods, such as low accuracy and poor interpretability in medical practice. This framework integrates causal discovery algorithms, clinical insights, and medical expertise to identify the key relationships between machine learning models and medical principles, ultimately improving diagnostic accuracy. Finally, the proposed system achieves an 83.3% success rate in distinguishing between healthy, sub-healthy, and unhealthy knee states, outperforming traditional methods with an accuracy improvement of 6.4% to 17%. Furthermore, the system integrates a wireless sensor network and user-friendly software, enabling convenient monitoring and diagnosis in both clinical and home settings. The main innovation of this study lies in the design of a wearable sensor network structure based on patellar activity and a medical knowledge-guided Bayesian framework to enable a three-state classification diagnosis, namely, healthy, sub-healthy, and unhealthy. This system is particularly suited for knee health assessments in outpatient clinics, community settings, and home environments..

Glioblastoma (GBM) is a highly invasive brain tumor, whose cells infiltrate surrounding normal brain tissue beyond the lesion outlines visible in the current medical scans. These infiltrative cells are treated mainly by radiotherapy. Existing radiotherapy plans for brain tumors derive from population studies and scarcely account for patient-specific conditions. Here, we provide a Bayesian machine learning framework for the rational design of improved, personalized radiotherapy plans using mathematical modeling and patient multimodal medical scans. Our method, for the first time, integrates complementary information from high-resolution MRI scans and highly specific FET-PET metabolic maps to infer tumor cell density in GBM patients. The Bayesian framework quantifies imaging and modeling uncertainties and predicts patient-specific tumor cell density with credible intervals. The proposed methodology relies only on data acquired at a single time point and, thus, is applicable to standard clinical settings. An initial clinical population study shows that the radiotherapy plans generated from the inferred tumor cell infiltration maps spare more healthy tissue thereby reducing radiation toxicity while yielding comparable accuracy with standard radiotherapy protocols. Moreover, the inferred regions of high tumor cell densities coincide with the tumor radioresistant areas, providing guidance for personalized dose-escalation. The proposed integration of multimodal scans and mathematical modeling provides a robust, non-invasive tool to assist personalized radiotherapy design.

In this article we study the sparse signal recovery problem in a Bayesian framework using a novel Bootstrapped Sparse Bayesian Learning method. Sparse Bayesian Learning (SBL) framework is an effective tool for pruning out the irrelevant features and ending up with a sparse representation. In SBL the choice of prior over the variances of the Gaussian Scale mixture has been an interesting area of research for some time now. This motivates us to use a more generalized maximum entropy density as the prior which results in a new variant of SBL. It has been shown to perform better than traditional SBL empirically and it also accelerates the pruning procedure. Because of this advantage, this variant of SBL can be claimed as more robust choice as it is less sensitive to the threshold for pruning. Theoretical justifications have also been provided to show that the proposed model actually promotes sparse point estimates.

Nondiffusive phonon transport presents significant challenges in micro/nanoscale thermal characterization, compounded by the limitations of experimental-numerical techniques and the presence of measurement noise. Additionally, inverse modeling and uncertainty quantification (UQ) for submicron thermal transport remain under-explored. In this study, we introduce a physics-informed Bayesian deep learning framework designed to address phonon Boltzmann transport equation (BTE)-based forward and inverse problems leveraging limited and noisy data. Our approach combines Bayesian neural networks with a nonparametric variational inference method, formulating the BTE-constrained training in a Bayesian manner. This enables the estimation of the posterior distribution of neural network parameters and unknown equation parameters based on a likelihood function that incorporates uncertainties from both the measurement data and the BTE model. Through numerical experiments on various phonon transport scenarios, we demonstrate that our method can accurately reconstruct temperature and heat flux profiles, infer critical quantities of interest (e.g., Knudsen number), and provide robust uncertainty quantification, even when data is sparse and noisy. This framework enhances our capability to conduct nondiffusive thermal simulations and inverse modeling with quantified uncertainty, offering a powerful tool for advancing thermal transport research and optimization in micro/nanoscale devices.