RESUMEN

In the realm of ophthalmic surgeries, silicone oil is often utilized as a tamponade agent for repairing retinal detachments, but it necessitates subsequent removal. This study harnesses the power of machine learning to analyze the macular and optic disc perfusion changes pre and post-silicone oil removal, using Optical Coherence Tomography Angiography (OCTA) data. Building upon the foundational work of prior research, our investigation employs Gaussian Process Regression (GPR) and Long Short-Term Memory (LSTM) networks to create predictive models based on OCTA scans. We conducted a comparative analysis focusing on the flow in the outer retina and vessel density in the deep capillary plexus (superior-hemi and perifovea) to track perfusion changes across different time points. Our findings indicate that while machine learning models predict the flow in the outer retina with reasonable accuracy, predicting the vessel density in the deep capillary plexus (particularly in the superior-hemi and perifovea regions) remains challenging. These results underscore the potential of machine learning to contribute to personalized patient care in ophthalmology, despite the inherent complexities in predicting ocular perfusion changes.

Asunto(s)

RESUMEN

Suspensions containing microencapsulated phase change materials (MPCMs) play a crucial role in thermal energy storage (TES) systems and have applications in building materials, textiles, and cooling systems. This study focuses on accurately predicting the dynamic viscosity, a critical thermophysical property, of suspensions containing MPCMs and MXene particles using Gaussian process regression (GPR). Twelve hyperparameters (HPs) of GPR are analyzed separately and classified into three groups based on their importance. Three metaheuristic algorithms, namely genetic algorithm (GA), particle swarm optimization (PSO), and marine predators algorithm (MPA), are employed to optimize HPs. Optimizing the four most significant hyperparameters (covariance function, basis function, standardization, and sigma) within the first group using any of the three metaheuristic algorithms resulted in excellent outcomes. All algorithms achieved a reasonable R-value (0.9983), demonstrating their effectiveness in this context. The second group explored the impact of including additional, moderate-significant HPs, such as the fit method, predict method and optimizer. While the resulting models showed some improvement over the first group, the PSO-based model within this group exhibited the most noteworthy enhancement, achieving a higher R-value (0.99834). Finally, the third group was analyzed to examine the potential interactions between all twelve HPs. This comprehensive approach, employing the GA, yielded an optimized GPR model with the highest level of target compliance, reflected by an impressive R-value of 0.999224. The developed models are a cost-effective and efficient solution to reduce laboratory costs for various systems, from TES to thermal management.

RESUMEN

Modern machine learning has the potential to fundamentally change the way bioprocesses are developed. In particular, horizontal knowledge transfer methods, which seek to exploit data from historical processes to facilitate process development for a new product, provide an opportunity to rethink current workflows. In this work, we first assess the potential of two knowledge transfer approaches, meta learning and one-hot encoding, in combination with Gaussian process (GP) models. We compare their performance with GPs trained only on data of the new process, that is, local models. Using simulated mammalian cell culture data, we observe that both knowledge transfer approaches exhibit test set errors that are approximately halved compared to those of the local models when two, four, or eight experiments of the new product are used for training. Subsequently, we address the question whether experiments for a new product could be designed more effectively by exploiting existing knowledge. In particular, we suggest to specifically design a few runs for the novel product to calibrate knowledge transfer models, a task that we coin calibration design. We propose a customized objective function to identify a set of calibration design runs, which exploits differences in the process evolution of historical products. In two simulated case studies, we observed that training with calibration designs yields similar test set errors compared to common design of experiments approaches. However, the former requires approximately four times fewer experiments. Overall, the results suggest that process development could be significantly streamlined when systematically carrying knowledge from one product to the next.

Asunto(s)

RESUMEN

Accurately estimating Battery State of Charge (SOC) is essential for safe and optimal electric vehicle operation. This paper presents a comparative assessment of multiple machine learning regression algorithms including Support Vector Machine, Neural Network, Ensemble Method, and Gaussian Process Regression for modelling the complex relationship between real-time driving data and battery SOC. The models are trained and tested on extensive field data collected from diverse drivers across varying conditions. Statistical performance metrics evaluate the SOC prediction accuracy on the test set. Gaussian process regression demonstrates superior precision surpassing the other techniques with the lowest errors. Case studies analyse model competence in mimicking actual battery charge/discharge characteristics responding to changing drivers, temperatures, and drive cycles. The research provides a reliable data-driven framework leveraging advanced analytics for precise real-time SOC monitoring to enhance battery management.

RESUMEN

This study explores machine learning (ML) capabilities for predicting the shear strength of reinforced concrete deep beams (RCDBs). For this purpose, eight typical machine-learning models, i.e., symbolic regression (SR), XGBoost (XGB), CatBoost (CATB), random forest (RF), LightGBM, support vector regression (SVR), artificial neural networks (ANN), and Gaussian process regression (GPR) models, are selected and compared based on a database of 840 samples with 14 input features. The hyperparameter tuning of the introduced ML models is performed using the Bayesian optimization (BO) technique. The comparison results show that the CatBoost model is the most reliable and accurate ML model (R2 = 0.997 and 0.947 in the training and testing sets, respectively). In addition, simple and practical design expressions for RCDBs have been proposed based on the SR model with a physical meaning and acceptable accuracy (an average prediction-to-test ratio of 0.935 and a standard deviation of 0.198). Meanwhile, the shear strength predicted by ML models was then compared with classical mechanics-driven shear models, including two prominent practice codes (i.e., ACI318, EC2) and two previous mechanical models, which indicated that the ML approach is highly reliable and accurate over conventional methods. In addition, a reliability-based design was conducted on two ML models, and their reliability results were compared with those of two code standards. The findings revealed that the ML models demonstrate higher reliability compared to code standards.

RESUMEN

The advent of digital twins facilitates the generation of high-fidelity replicas of actual systems or assets, thereby enhancing the design's performance and feasibility. When developing digital twins, precise measurement data is essential to ensure alignment between the actual and digital models. However, inherent uncertainties in sensors and models lead to disparities between observed and predicted (simulated) behaviors. To mitigate these uncertainties, this study originally proposes a multi-objective optimization strategy utilizing a Gaussian process regression surrogate model, which integrates various uncertain parameters, such as load angle, bucket cylinder stroke, arm cylinder stroke, and boom cylinder stroke. This optimization employs a genetic algorithm to indicate the Pareto frontiers regarding the pressure exerted on the boom, arm, and bucket cylinders. Subsequently, TOPSIS is applied to ascertain the optimal candidate among the identified Pareto optima. The findings reveal a substantial congruence between the experimental and numerical outcomes of the devised virtual model, in conjunction with the TOPSIS-derived optimal parameter configuration.

RESUMEN

Street cocaine is often mixed with various substances that intensify its harmful effects. This paper proposes a framework to identify attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) intervals that best predict the concentration of adulterants in cocaine samples. Wavelengths are ranked according to their relevance through ReliefF and mRMR feature selection approaches, and an iterative process removes less relevant wavelengths based on the ranking suggested by each approach. Gaussian Process (GP) regression models are constructed after each wavelength removal and the prediction performance is evaluated using RMSE. The subset balancing a low RMSE value and a small percentage of retained wavelengths is chosen. The proposed framework was validated using a dataset consisting of 345 samples of cocaine with different amounts of levamisole, caffeine, phenacetin, and lidocaine. Averaged over the four adulterants, the GP regression coupled with the mRMR retained 1.07 % of the 662 original wavelengths, outperforming PLS and SVR regarding prediction performance.

Asunto(s)

RESUMEN

This study presents a new approach to determining the preload force of bolted joints. The concept involves measuring the torsional angle without contact. For this purpose, we present a circular magnetic sensor array integrated into the torque wrench. The torsional angle in bolted joints depends on the dimensions of the screw and the materials used and is typically less than four degrees. For this reason, one requirement is a high angular resolution so that a continuous recording of the torsion angle is feasible during the assembly process. This can be achieved using the circular sensor array and adapted signal processing methods. Two signal processing approaches are utilized. First, the direct method uses the discrete Fourier transformation to calculate the rotation angle from the signal phase. This approach is robust to signal distortion and does not depend on signal amplitude. Second, the method with a learning phase employs Gaussian process regression to minimize the angle error. In an experiment, both approaches were applied within a test bench and showed promising results. The direct method demonstrated a very good angular resolution without training and calibration. For mobile and less-complex applications where a reference system is unavailable, the direct method is preferable. However, in complex measurement systems where reference systems can be utilized initially, significant enhancements to an excellent resolution can be achieved through prior training.

RESUMEN

In recent times, many investigators have delved into plastic waste (PW) research, both locally and internationally. Many of these studies have focused on problems related to land-based and marine-based PW management with its attendant impact on public health and the ecosystem. Hitherto, there have been little or no studies on forecasting PW quantities in developing countries (DCs). The key objective of this study is to provide a forecast on PW generation in the city of Johannesburg (CoJ), South Africa over the next three decades. The data used for the forecasting were historical data obtained from Statistics South Africa (StatsSA). For effective prediction and comparison, three-time series models were employed in this study. They include exponential smoothing (ETS), Artificial Neural Network (ANN), and the Gaussian Process Regression (GPR). The exponential kernel GPR model performed best on the overall plastic prediction with a determination coefficient (R2) of 0.96, however, on individual PW estimation, ANN was better with an overall R2 of 0.93. From the result, it is predicted that between 2021 and 2050, the total PW generated in CoJ is forecasted to be around 6.7 megatonnes with an average of 0.22 megatonnes/year. In addition, the estimated plastic composition is 17,910 tonnes PS per year; 13,433 tonnes PP per year; 59,440 tonnes HDPE per year; 4478 tonnes PVC per year; 85,074 tonnes PET per year; 34,590 tonnes LDPE per year and 8955 tonnes other PWs per year.

RESUMEN

INTRODUCTION: Nonstationary, nonlinear mass transfer in traditional Chinese medicine (TCM) extraction poses challenges to correlating process characteristics with quality parameters, particularly in defining clear parameter ranges for the process. OBJECTIVES: The aim of the study was to provide a solution for quality consistency analysis in TCM preparation processes. MATERIALS AND METHODS: Salvia miltiorrhiza was taken as an example for 15 batches of standard decoction. Using aqueous extract, alcoholic extract, and the content of salvianolic acid B as herb material key quality attributes, multiple nonlinear regression, Gaussian process regression, and artificial neural network models were employed to predict the key quality attributes including the paste yield, the content of salvianolic acid B, and the transfer rate. The evaluation criteria were root mean square error, mean absolute percentage error, and R2. RESULTS: The Gaussian process regression model had the best prediction effect on the paste yield, the content of salvianolic acid B, and the transfer rate, with R2 being 0.918, 0.934, and 0.919, respectively. Utilizing Gaussian process regression model confidence intervals, along with Shewhart control and intervals optimized through process capability index analysis, the quality control range of the standard decoction was determined as follows: paste yield, 25.14%-33.19%; salvianolic acid B content, 2.62%-4.78%; and transfer rate, 56.88%-64.80%. CONCLUSION: This study combined the preparation process of standard decoction with the Gaussian process regression model, accurately predicted the key quality attributes, and determined the quality parameter range by using process analysis tools, providing a new idea for the quality consistency standard of TCM processes.

Asunto(s)

RESUMEN

In addition to the filter coefficients, the location of the microphone array is a crucial factor in improving the overall performance of a beamformer. The optimal microphone array placement can considerably enhance speech quality. However, the optimization problem with microphone configuration variables is non-convex and highly non-linear. Heuristic algorithms that are frequently employed take a long time and have a chance of missing the optimal microphone array placement design. We extend the Bayesian optimization method to solve the microphone array configuration design problem. The proposed Bayesian optimization method does not depend on gradient and Hessian approximations and makes use of all the information available from prior evaluations. Furthermore, Gaussian process regression and acquisition functions make up the Bayesian optimization method. The objective function is given a prior probabilistic model through Gaussian process regression, which exploits this model while integrating out uncertainty. The acquisition function is adopted to decide the next placement point based upon the incumbent optimum with the posterior distribution. Numerical experiments have demonstrated that the Bayesian optimization method could find a similar or better microphone array placement compared with the hybrid descent method and computational time is significantly reduced. Our proposed method is at least four times faster than the hybrid descent method to find the optimal microphone array configuration from the numerical results.

RESUMEN

Surface roughness prediction is a pivotal aspect of the manufacturing industry, as it directly influences product quality and process optimization. This study introduces a predictive model for surface roughness in the turning of complex-structured workpieces utilizing Gaussian Process Regression (GPR) informed by vibration signals. The model captures parameters from both the time and frequency domains of the turning tool, encompassing the mean, median, standard deviation (STD), and root mean square (RMS) values. The signal is from the time to frequency domain and it is executed using Welch's method complemented by time-frequency domain analysis employing three levels of Daubechies Wavelet Packet Transform (WPT). The selected features are then utilized as inputs for the GPR model to forecast surface roughness. Empirical evidence indicates that the GPR model can accurately predict the surface roughness of turned complex-structured workpieces. This predictive strategy has the potential to improve product quality, streamline manufacturing processes, and minimize waste within the industry.

RESUMEN

This study explores the potential of photocatalytic degradation using novel NML-BiFeO3 (noble metal-incorporated bismuth ferrite) compounds for eliminating malachite green (MG) dye from wastewater. The effectiveness of various Gaussian process regression (GPR) models in predicting MG degradation is investigated. Four GPR models (Matern, Exponential, Squared Exponential, and Rational Quadratic) were employed to analyze a dataset of 1200 observations encompassing various experimental conditions. The models have considered ten input variables, including catalyst properties, solution characteristics, and operational parameters. The Exponential kernel-based GPR model achieved the best performance, with a near-perfect R2 value of 1.0, indicating exceptional accuracy in predicting MG degradation. Sensitivity analysis revealed process time as the most critical factor influencing MG degradation, followed by pore volume, catalyst loading, light intensity, catalyst type, pH, anion type, surface area, and humic acid concentration. This highlights the complex interplay between these factors in the degradation process. The reliability of the models was confirmed by outlier detection using William's plot, demonstrating a minimal number of outliers (66-71 data points depending on the model). This indicates the robustness of the data utilized for model development. This study suggests that NML-BiFeO3 composites hold promise for wastewater treatment and that GPR models, particularly Matern-GPR, offer a powerful tool for predicting MG degradation. Identifying fundamental catalyst properties can expedite the application of NML-BiFeO3, leading to optimized wastewater treatment processes. Overall, this study provides valuable insights into using NML-BiFeO3 compounds and machine learning for efficient MG removal from wastewater.

RESUMEN

Radioactive wastes contain organic complexing agents that can form complexes with radionuclides and enhance the solubility of these radionuclides, increasing the mobility of radionuclides over great distances from a radioactive waste repository. In this study, four radionuclides (cobalt, strontium, iodine, and uranium) and three organic complexing agents (ethylenediaminetetraacetic acid, nitrilotriacetic acid, and iso-saccharic acid) were selected, and the solubility of these radionuclides was assessed under realistic environmental conditions such as different pHs (7, 9, 11, and 13), temperatures (10 °C, 20 °C, and 40 °C), and organic complexing agent concentrations (10-5-10-2 M). A total of 720 datasets were generated from solubility batch experiments. Four supervised machine learning models such as the Gaussian process regression (GPR), ensemble-boosted trees, artificial neural networks, and support vector machine were developed for predicting the radionuclide solubility. Each ML model was optimized using Bayesian optimization algorithm. The GPR evolved as a robust model that provided accurate predictions within the underlying solubility patterns by capturing the intricate relationships of the independent parameters of the dataset. At an uncertainty level of 95%, both the experimental results and GPR simulated estimations were closely correlated, confirming the suitability of the GPR model for future explorations.

RESUMEN

Machine learning (ML) force fields are revolutionizing molecular dynamics (MD) simulations as they bypass the computational cost associated with ab initio methods but do not sacrifice accuracy in the process. In this work, the GPyTorch library is used to create Gaussian process regression (GPR) models that are interfaced with the next-generation ML force field FFLUX. These models predict atomic properties of different molecular configurations that appear in a progressing MD simulation. An improved kernel function is utilized to correctly capture the periodicity of the input descriptors. The first FFLUX molecular simulations of ammonia, methanol, and malondialdehyde with the updated kernel are performed. Geometry optimizations with the GPR models result in highly accurate final structures with a maximum root-mean-squared deviation of 0.064 Å and sub-kJ mol-1 total energy predictions. Additionally, the models are tested in 298 K MD simulations with FFLUX to benchmark for robustness. The resulting energy and force predictions throughout the simulation are in excellent agreement with ab initio data for ammonia and methanol but decrease in quality for malondialdehyde due to the increased system complexity. GPR model improvements are discussed, which will ensure the future scalability to larger systems.

RESUMEN

The integration time and signal-to-noise ratio are inextricably linked when performing scanning probe microscopy based on raster scanning. This often yields a large lower bound on the measurement time, for example, in nano-optical imaging experiments performed using a scanning near-field optical microscope (SNOM). Here, we utilize sparse scanning augmented with Gaussian process regression to bypass the time constraint. We apply this approach to image charge-transfer polaritons in graphene residing on ruthenium trichloride (α-RuCl3) and obtain key features such as polariton damping and dispersion. Critically, nano-optical SNOM imaging data obtained via sparse sampling are in good agreement with those extracted from traditional raster scans but require 11 times fewer sampled points. As a result, Gaussian process-aided sparse spiral scans offer a major decrease in scanning time.

RESUMEN

Since the introduction of the BrainAGE method, novel machine learning methods for brain age prediction have continued to emerge. The idea of estimating the chronological age from magnetic resonance images proved to be an interesting field of research due to the relative simplicity of its interpretation and its potential use as a biomarker of brain health. We revised our previous BrainAGE approach, originally utilising relevance vector regression (RVR), and substituted it with Gaussian process regression (GPR), which enables more stable processing of larger datasets, such as the UK Biobank (UKB). In addition, we extended the global BrainAGE approach to regional BrainAGE, providing spatially specific scores for five brain lobes per hemisphere. We tested the performance of the new algorithms under several different conditions and investigated their validity on the ADNI and schizophrenia samples, as well as on a synthetic dataset of neocortical thinning. The results show an improved performance of the reframed global model on the UKB sample with a mean absolute error (MAE) of less than 2 years and a significant difference in BrainAGE between healthy participants and patients with Alzheimer's disease and schizophrenia. Moreover, the workings of the algorithm show meaningful effects for a simulated neocortical atrophy dataset. The regional BrainAGE model performed well on two clinical samples, showing disease-specific patterns for different levels of impairment. The results demonstrate that the new improved algorithms provide reliable and valid brain age estimations.

Asunto(s)

RESUMEN

Personalized treatment informed by computational models has the potential to markedly improve the outcome for patients with a type B aortic dissection. However, existing computational models of dissected walls significantly simplify the characteristic false lumen, tears and/or material behavior. Moreover, the patient-specific wall thickness and stiffness cannot be accurately captured non-invasively in clinical practice, which inevitably leads to assumptions in these wall models. It is important to evaluate the impact of the corresponding uncertainty on the predicted wall deformations and stress, which are both key outcome indicators for treatment optimization. Therefore, a physiology-inspired finite element framework was proposed to model the wall deformation and stress of a type B aortic dissection at diastolic and systolic pressure. Based on this framework, 300 finite element analyses, sampled with a Latin hypercube, were performed to assess the global uncertainty, introduced by 4 uncertain wall thickness and stiffness input parameters, on 4 displacement and stress output parameters. The specific impact of each input parameter was estimated using Gaussian process regression, as surrogate model of the finite element framework, and a δ moment-independent analysis. The global uncertainty analysis indicated minor differences between the uncertainty at diastolic and systolic pressure. For all output parameters, the 4th quartile contained the major fraction of the uncertainty. The parameter-specific uncertainty analysis elucidated that the material stiffness and relative thickness of the dissected membrane were the respective main determinants of the wall deformation and stress. The uncertainty analysis provides insight into the effect of uncertain wall thickness and stiffness parameters on the predicted deformation and stress. Moreover, it emphasizes the need for probabilistic rather than deterministic predictions for clinical decision making in aortic dissections.

Asunto(s)

RESUMEN

Electron holography provides quantitative phase information regarding the electromagnetic fields and the morphology of micro- to nano-scale samples. A phase image reconstructed numerically from an electron hologram sometimes includes phase residues, i.e. origins of unremovable phase discontinuities, which make it much more difficult to quantitatively analyze local phase values. We developed a method to remove the residues in a phase image by a combination of patching local areas of a hologram and denoising based on machine learning. The small patches for a hologram, which were generated using the spatial frequency information of the own fringe patterns, were pasted at each residue point by an algorithm based on sparse modeling. After successive phase reconstruction, the phase components with no dependency on the vicinity were filtered out by Gaussian process regression. We determined that the phase discontinuities that appeared around phase residues were removed and the phase distributions of an atomic resolution phase image of a Pt nanoparticle were sufficiently restored.

RESUMEN

In the last several years, there has been a surge in the development of machine learning potential (MLP) models for describing molecular systems. We are interested in a particular area of this field - the training of system-specific MLPs for reactive systems - with the goal of using these MLPs to accelerate free energy simulations of chemical and enzyme reactions. To help new members in our labs become familiar with the basic techniques, we have put together a self-guided Colab tutorial (https://cc-ats.github.io/mlp_tutorial/), which we expect to be also useful to other young researchers in the community. Our tutorial begins with the introduction of simple feedforward neural network (FNN) and kernel-based (using Gaussian process regression, GPR) models by fitting the two-dimensional Müller-Brown potential. Subsequently, two simple descriptors are presented for extracting features of molecular systems: symmetry functions (including the ANI variant) and embedding neural networks (such as DeepPot-SE). Lastly, these features will be fed into FNN and GPR models to reproduce the energies and forces for the molecular configurations in a Claisen rearrangement reaction.