Synthetic Samples Research Articles

AI Based Voice Cloning and Generation for Vocally Challenged

Voice cloning, the ability to replicate a person’s voice with high fidelity, has significant applications in entertainment, accessibility, virtual assistants, and forensic sciences. Recent advancements in deep learning have demonstrated the potential of Generative Adversarial Networks (GANs) to generate realistic synthetic audio. This study explores the application of GANs for voice cloning, leveraging their adversarial training mechanism to achieve high-quality and natural-sounding voice synthesis. The proposed framework integrates a generator network designed to produce realistic audio waveforms and a discriminator network tasked with distinguishing between real and synthetic samples. The system is trained on a dataset of diverse voice recordings, focusing on capturing both prosody and speaker-specific features. To enhance the cloning process, the model employs auxiliary losses, such as mel-spectrogram reconstruction and perceptual loss, ensuring the generated audio aligns closely with human perception. Experimental results demonstrate that GAN-based voice cloning outperforms traditional methods in both audio quality and speaker similarity, even with limited data. The research also discusses ethical considerations, including misuse risks and countermeasures, emphasizing the importance of responsible deployment. The findings establish GANs as a promising approach for advancing voice cloning technology while highlighting avenues for future research in robustness and real-time applications.

Experimental Study: Stress Path Coefficient in Unconsolidated Sands: Effects of Re-Pressurization and Depletion Hysteresis

Accurate estimation of in-situ stresses is a critical parameter for geo-mechanical modelling. In-situ stresses are estimated in the field from logs and frac tests. Laboratory tests are performed with cored material to estimate horizontal stress changes under defined boundary conditions to complement field data. Horizontal stress path coefficient is used to estimate a change in in-situ stresses as the reservoir undergoes depletion or injection. Uniaxial Strain boundary conditions are representative of far field stress state. The laboratory data provides the change in horizontal stress with a change in pore pressure. It is used to complement the field data acquisition of absolute stress values to predict the value of total stresses. This experimental study provides a novel method of simulating geological compaction for fabricating representative samples from unconsolidated sands. It investigates the variability of horizontal stress path coefficient as a function of changing pore pressure (depressurization and re-pressurization) in unconsolidated sandstone reservoirs. Synthetic sandstones samples were made from sand packs by consolidating them under an isostatic stress path at ambient pore pressure. After getting to initial reservoir conditions, a series of pore pressure depletion and injection tests with varying magnitudes (injection and depletion) were performed to study the effects of stress path direction and associated hysteresis. The magnitude of the stress path coefficient under depletion is lower than that under injection for the first load-unload cycle. In subsequent load-unload cycles, the stress path coefficient values remain constant until the sample is depleted to a new level of pore pressure. A Modified Cam Clay model is fit to the data to map the expansion of the yield surface and quantify the model parameters. Application of this research includes accurate prediction of changes in-situ stresses during depletion and injection stress paths for simulating unconsolidated reservoirs behavior under fluid injection or further depletion.

Future-proofing class-incremental learning

Exemplar-free class incremental learning is a highly challenging setting where replay memory is unavailable. Methods relying on frozen feature extractors have drawn attention recently in this setting due to their impressive performances and lower computational costs. However, those methods are highly dependent on the data used to train the feature extractor and may struggle when an insufficient amount of classes are available during the first incremental step. To overcome this limitation, we propose to use a pre-trained text-to-image diffusion model in order to generate synthetic images of future classes and use them to train the feature extractor. Experiments on the standard benchmarks CIFAR100 and ImageNet-Subset demonstrate that our proposed method can be used to improve state-of-the-art methods for exemplar-free class incremental learning, especially in the most difficult settings where the first incremental step only contains few classes. Moreover, we show that using synthetic samples of future classes achieves higher performance than using real data from different classes, paving the way for better and less costly pre-training methods for incremental learning.

Retinal Image Augmentation using Composed GANs

Medical image analysis faces a significant challenge in the scarcity of annotated data, which is crucial for developing generalizable Deep Learning (DL) models that require extensive training data. Consequently, the field of medical image generation has garnered substantial interest and potential for further exploration. Besides widely employed data augmentation techniques, such as rotation, reflection, and scaling, Generative Adversarial Networks (GANs) have demonstrated the ability to effectively leverage additional information from datasets by generating synthetic samples from real images. In the context of retinal image synthesis, an image-to-image translation approach is frequently adopted to generate retinal images from available vessel maps, which can be scarce and resource-intensive to obtain. Deviating from prior work reliant on pre-existing vessel maps, this study proposes a learning-based model that is independent of vessel maps, utilizing Progressive Growing GAN (PGGAN) to generate vascular networks from random noise. The visual and quantitative evaluations conducted suggest that the majority of the images generated by the proposed model are substantially distinct from the training set while maintaining a high proportion of true image quality, underscoring the model's potential as a powerful tool for data augmentation.

Deep-learning and data-resampling: A novel approach to predict cyanobacterial alert levels in a reservoir

The proliferation of harmful algal blooms results in adverse impacts on aquatic ecosystems and public health. Early warning system monitors algal bloom occurrences and provides management strategies for promptly addressing high-concentration algal blooms following their occurrence. In this study, we aimed to develop a proactive prediction model for cyanobacterial alert levels to enable efficient decision-making in management practices. We utilized 11 years of water quality, hydrodynamic, and meteorological data from a reservoir that experiences frequent harmful cyanobacterial blooms in summer. We used these data to construct a deep-learning model, specifically a 1D convolution neural network (1D-CNN) model, to predict cyanobacterial alert levels one week in advance. However, the collected distribution of algal alert levels was imbalanced, leading to the biased training of data-driven models and performance degradation in model predictions. Therefore, an adaptive synthetic sampling method was applied to address the imbalance in the minority class data and improve the predictive performance of the 1D-CNN. The adaptive synthetic sampling method resolved the imbalance in the data during the training phase by incorporating an additional 156 and 196 data points for the caution and warning levels, respectively. The selected optimal 1D-CNN model with a filter size of 5 and comprising 16 filters achieved training and testing prediction accuracies of 97.3% and 85.0%, respectively. During the test phase, the prediction accuracies for each algal alert level (L-0, L-1, and L-2) were 89.9%, 79.2%, and 71.4%, respectively, indicating reasonably consistent predictive results for all three alert levels. Therefore, the use of synthetic data addressed data imbalances and enhanced the predictive performance of the data-driven model. The reliable forecasts produced by the improved model can support the development of management strategies to mitigate harmful algal blooms in reservoirs and can aid in building an early warning system to facilitate effective responses.

Measurement Guidance in Diffusion Models: Insight from Medical Image Synthesis.

In the field of healthcare, the acquisition of sample is usually restricted by multiple considerations, including cost, labor- intensive annotation, privacy concerns, and radiation hazards, therefore, synthesizing images-of-interest is an important tool to data augmentation. Diffusion models have recently attained state-of-the-art results in various synthesis tasks, and embedding energy functions has been proved that can effectively guide the pre-trained model to synthesize target samples. However, we notice that current method development and validation are still limited to improving indicators, such as Fréchet Inception Distance score (FID) and Inception Score (IS), and have not provided deeper investigations on downstream tasks, like disease grading and diagnosis. Moreover, existing classifier guidance which can be regarded as a special case of energy function can only has a singular effect on altering the distribution of the synthetic dataset. This may contribute to in-distribution synthetic sample that has limited help to downstream model optimization. All these limitations remind that we still have a long way to go to achieve controllable generation. In this work, we first conducted an analysis on previous guidance as well as its contributions on further applications from the perspective of data distribution. To synthesize samples which can help downstream applications, we then introduce uncertainty guidance in each sampling step and design an uncertainty-guided diffusion models. Extensive experiments on four medical datasets, with ten classic networks trained on the augmented sample sets provided a comprehensive evaluation on the practical contributions of our methodology. Furthermore, we provide a theoretical guarantee for general gradient guidance in diffusion models, which would benefit future research on investigating other forms of measurement guidance for specific generative tasks.

Using machine learning to reveal seasonal nutrient dynamics and their impact on chlorophyll-a levels in lake ecosystems: A focus on nitrogen and phosphorus

Chlorophyll-a (Chl-a) is a pivotal indicator of lake eutrophication. Studies examining nutrients limiting lake eutrophication at large scales have traditionally focused on summer and autumn, potentially limiting the applicability of their findings. This study encompasses 86 state-controlled points in the Eastern China Basin, spanning data collected from January 2020 to July 2023. Furthermore, we focus on the application of three machine-learning models (i.e., eXtreme Gradient Boosting, Support Vector Machines, and Naive Bayes Classifier) to analyze the seasonal nutrient dynamics in lake ecosystems. We categorized the monitoring data by season to eliminate outliers and employed adaptive synthetic sampling to address data imbalance issues. The results reveal that the direct correlations between total nitrogen (TN), total phosphorus (TP), and TP in conjunction with turbidity and Chl-a are broadly weak, possibly because of geographic variations, nutrient lag effects on algae, and differences in algal community composition. However, probabilistic analyses revealed that as TP or TN levels transitioned from oligo-mesotrophic (O) to eutrophic (E), TP exhibited a greater influence on the variation in Chl-a status than TN during spring and winter (p < 0.05). Conversely, the effects of TP and TN on Chl-a (O-E) were comparable during summer and autumn. Seasonal variations in TN and TP thresholds derived from XGBoost modeling for O and E states of Chl-a suggest the need for stricter control measures during periods of high-nutrient levels and cost-effective management strategies to employ during low-nutrient periods. These findings should enhance our understanding of trophic shifts in lakes and provide a foundation for optimizing eutrophication management strategies across all seasons.

Farmers' credit risk evaluation with an explainable hybrid ensemble approach: A closer look in microfinance

Artificial intelligence stimulates the vitality of microcredit by reshaping credit risk evaluation models, especially targeting the group of farmers. Therefore, the paper aims to establish a new interpretable hybrid ensemble model for evaluating the credit risk of microfinance for farmers, which is called ADASYN (Adaptive Synthetic Sampling)-LCE (Local Cascade Ensemble)-Shapash. It integrates the advantages of three ensemble models: bagging, boosting, and local cascading, including reducing model variance, reducing model bias, and simplifying complex problems by learning different parts of the training data. And it alleviates the problem of low generalization performance of traditional ensemble models caused by imbalanced loan data of farmers. Through the empirical analysis of the data of farmers' loans of China poverty alleviation agency “CHONGHO BRIDGE”, it is found that its average rank is 2.1, which is better than other integrated models in the credit risk evaluation of farmers' microfinance. Finally, the global and local interpretation of our model is preliminarily explored.

Development of a green-synthesized molecularly imprinted polymer-based electrochemical nanosensor for the determination of N-nitrosodimethylamine (NDMA) in serum and tap water.

N-nitrosodimethylamine (NDMA) was determined using a molecularly imprinted polymer (MIP)-based electrochemical sensor.Green-synthesized silver nanoparticles were functionalized with cysteamine to enhance their integration into the electrode surface, which was used to modify a glassy carbon electrode (GCE). Furthermore, a MIP-based electrochemical sensor was constructed via electropolymerization of 3-aminophenyl boronic acid (3-APBA) as a conjugated functional monomer in the presence of lithium perchlorate (LiClO4) solution as a dopant, chitosan as a carrier natural polymer, and NDMA as a template/target molecule. The polymer film was characterized by scanning electron microscopy (SEM) and electrochemical impedance spectroscopy (EIS). The analytical performance of the silver nanomaterial-based MIP-based electrochemical (AgNPs@Chitosan/3-APBA@MIP-GCE) sensor was evaluated under optimized conditions. The linear range of NDMA was 1.0 × 10-13-1.0 × 10-12M (0.1-1.0pM), with a limit of detection (LOD) of 3.63 × 10-15M (3.63 fM) using differential pulse voltammetry (DPV). Method validation figured out that the developed MIP-based electrochemical nanosensor exhibited excellent selectivity, accuracy, and precision, which was shown by the analysis of synthetic serum samples and tap water. The LOD and LOQ in serum samples were 17.8 fM and 59.5 fM,respectively,which were in agreement with the developed method. Good recovery results confirm the successful application of the method in serum and tap water samples. The selectivity of the developed AgNPs@Chitosan/3-APBA@MIP-GCE sensor for NDMA was demonstrated in the presence of NDEA, sartans (valsartan, losartan, irbesartan, candesartan, telmisartan), and potential interferents that arepossibly present in biological fluids (dopamine, ascorbic acid, uric acid) besides ionic species (sodium, chloride, potassium, nitrate, magnesium, sulfate) and common analgesic paracetamol.

Revealing efficacy of AgCuFe2O4@GO/MnO2 in 3D electrochemical oxidation for ceftriaxone degradation in aqueous media: Optimization and mechanisms

Ceftriaxone (CEF), a broad-spectrum antibiotic with a long half-life, is extensively used for the curative purposes of many bacterial infections. Nevertheless, given its strong resistance and the inefficiency of traditional techniques in breaking it down and eliminating it, this study assessed the use of a synthesized AgCuFe2O4@GO/MnO2 nanoparticle electrode (NPE) in a three-dimensional electrochemical oxidation reaction (3DER) to remove CEF. The AgCuFe2O4@GO/MnO2 NPE was fabricated by a co-precipitation process aided by a microwave. The physical and chemical structure of the nanocomposite was determined and verified using a range of analytical techniques, such as FESEM, XRD, EDS-mapping, FTIR, BET, VSM, and TGA. These investigations indicated that the NPE had a large specific surface area, a maintained crystal structure, strong magnetic characteristics, and a quasi-spherical morphology. A 3DER with optimal parameters (pH 5, initial CEF concentration 20 mg/L, NPE dose 0.7 g/L, electrode distance 3 cm, 0.12 mM persulfate electrolyte, and 8.5 mA/cm2 current density for 45 minutes) removed 86.8 % of CEF in synthetic samples and 71.3 % in real wastewater samples, with a mineralization rate of 53.4 %, also had 253.2 kWh/g energy consumption. The 3DER matched the pseudo-first-order kinetic and the Langmuir-Hinshelwood model (R2 > 0.9), with KC and KL-H values of 0.954 mg/L.min and 0.032 L/mg. The removal effectiveness of 64.9 % was achieved after five cycles of recovering and regenerating the NPE. The AgCuFe2O4@GO/MnO2 NPE is beneficial for treating a wide range of industrial and hospital wastewaters due to its magnetic characteristics, chemical stability, reusability, and remarkable efficiency.

Determination of 28 food safety risk factors in novel foods by ultra-high performance liquid chromatography-quadrupole/orbitrap high resolution mass spectrometry

Abstract An analytical method was established for the determination of 24 food safety risk factors (18 heterocyclic aromatic amines, 6 mycotoxins) in synthetic meat and synthetic protein samples, 4 microcystins in microalgae samples using ultra-high performance liquid chromatography-quadrupole/orbitrap high resolution mass spectrometry (UPLC-Q/Orbitrap HRMS). The novel synthetic food was extracted using a 50% methanol solution and purified with 3 mL of n-hexane saturated with a 50% methanol solution. For microalgae samples, extraction was performed with an 80% acetonitrile solution and purified using a PRiME HLB solid-phase extraction column. An ACQUITY HSS T3 column (100 mm ×2.1mm, 1.8 μm) was used for analyte separation, with 5 mmol/L ammonium acetate containing 0.01% formic acid and methanol selected as the mobile phase. Gradient elution was adopt. Data acquisition was conducted in both positive and negative mode using full scan-data dependent secondary scan mode (Full MS-ddMS2). Matrix-matched standards were employed for quantification in the external standard method. Additionally, mzVault and Tracefinder software were used to establish a screening database for 28 food safety risk factors. The results showed that the relative deviation of the exact mass of the 28 food safety risk factors was less than 5ppm, with good linearity in the range of 1-100μg/kg, and a correlation coefficient (r2) of ≥0.9978. The detection limit ranged from 0.5 to 4 μg/kg, and the quantification limit ranged from 1 to 10 μg/kg. Method recoveries ranged from 70.0% to 119.6% with relative standard deviations of 0.4% to 14.6% (n=6). The method is simple, sensitive and accurate, suitable for rapid screening of 28 food safety risk factors in novel foods.

UsfAD based effective unknown attack detection focused IDS framework

The rapid expansion of varied network systems, including the Internet of Things (IoT) and the Industrial Internet of Things (IIoT), has led to an increasing range of cyber threats. Ensuring robust protection against these threats necessitates the implementation of an effective Intrusion Detection System (IDS). For more than a decade, researchers have delved into supervised machine learning techniques to develop IDS to classify normal and attack traffic. However, building effective IDS models using supervised learning requires a substantial number of benign and attack samples. To collect a sufficient number of attack samples from real-life scenarios is not possible since cyber attacks occur occasionally. Further, IDS trained and tested on known datasets fails in detecting zero-day or unknown attacks due to the swift evolution of attack patterns. To address this challenge, we put forth two strategies for semi-supervised learning-based IDS where training samples of attacks are not required: (1) training a supervised machine learning model using randomly and uniformly dispersed synthetic attack samples; (2) building a One Class Classification (OCC) model that is trained exclusively on benign network traffic. We have implemented both approaches and compared their performances using 10 recent benchmark IDS datasets. Our findings demonstrate that the OCC model based on the state-of-art anomaly detection technique called usfAD significantly outperforms conventional supervised classification and other OCC-based techniques when trained and tested considering real-life scenarios, particularly to detect previously unseen attacks.

Enhancing crop yield prediction in Senegal using advanced machine learning techniques and synthetic data

In this study, we employ advanced data-driven techniques to investigate the complex relationships between the yields of five major crops and various geographical and spatiotemporal features in Senegal. We analyze how these features influence crop yields by utilizing remotely sensed data. Our methodology incorporates clustering algorithms and correlation matrix analysis to identify significant patterns and dependencies, offering a comprehensive understanding of the factors affecting agricultural productivity in Senegal. To optimize the model's performance and identify the optimal hyperparameters, we implemented a comprehensive grid search across four distinct machine learning regressors: Random Forest, Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and Light Gradient-Boosting Machine (LightGBM). Each regressor offers unique functionalities, enhancing our exploration of potential model configurations. The top-performing models were selected based on evaluating multiple performance metrics, ensuring robust and accurate predictive capabilities. The results demonstrated that XGBoost and CatBoost perform better than the other two. We introduce synthetic crop data generated using a Variational Auto Encoder to address the challenges posed by limited agricultural datasets. By achieving high similarity scores with real-world data, our synthetic samples enhance model robustness, mitigate overfitting, and provide a viable solution for small dataset issues in agriculture. Our approach distinguishes itself by creating a flexible model applicable to various crops together. By integrating five crop datasets and generating high-quality synthetic data, we improve model performance, reduce overfitting, and enhance realism. Our findings provide crucial insights for productivity drivers in key cropping systems, enabling robust recommendations and strengthening the decision-making capabilities of policymakers and farmers in data-scarce regions.

Near-Surface defect detection in ultrasonic testing using Domain-Knowledge-Informed self-supervised learning

Recently, significant research efforts have been made to enhance ultrasonic testing (UT) by employing artificial intelligence (AI). However, collecting an extensive amount of labeled data across various testing environments to train the AI model poses significant challenges. Moreover, conventional UT typically focuses on detecting deep-depth defects, which limits the effectiveness of such methods in detecting near-surface defects. To this end, this paper proposes a novel near-surface defect detection method for ultrasonic testing that can be employed without collecting labeled data. We propose a self-supervised anomaly detection model that incorporates domain knowledge. First, synthetic faulty samples are generated by fusing the measured UT signals with the back-wall UT reflection signals, to simulate real faulty features. Unlike the CutPaste method used for computer vision applications, this synthesis method adds the back-wall echo signal to random locations by incorporating the physical principles of the superposition of ultrasonic signals. Next, a de-anomaly network is devised to isolate subtle defect features within the measured UT signals. The presence of defects was determined using the three-sigma rule of the mean absolute value of the residual output. The defect depth is determined by a time-of-flight calculation from the residual output. The effectiveness of the proposed method was evaluated through the UT of aluminum blocks with near-surface defects of varying depths under different surface conditions. Both qualitative and quantitative comparison studies demonstrated that the proposed method outperformed existing methods in detecting the presence and depth of near-surface defects.

A hybrid machine learning approach for imbalanced irrigation water quality classification

Global food security is increasingly dependent on irrigation, particularly in regions experiencing freshwater scarcity. Conventional laboratory methods for assessing Irrigation Water Quality Index (IWQI) are often slow and inaccessible to small-scale farmers, especially in developing countries. This study proposes an efficient machine learning (ML) approach to enhance the classification performance of IWQI into five classes: no restriction, low restriction, moderate restriction, high restriction, and severe restriction. A dataset comprising 62,499 samples with six hydrochemical parameters (EC, Cl−, HCO3−, Na+, Ca2+, and Mg2+) was collected, preprocessed, and labeled. Missing values were imputed using a Random Forest model, achieving an R2 of 0.98. To address class imbalance, synthetic resampling, class weighting, and apportioned margins were employed to train three ML models: two stacked ensembles and an Apportioned Margin Support Vector Machine (AMSVM). Class weighting was applied to the first ensemble, adaptive synthetic sampling (ADASYN) resampling was utilized for the second, and AMSVM was adjusted for class imbalance through apportioned margins and class weighting. The class-weighted ensemble achieved 98.5% accuracy, precision, recall, and F1-score, while the ADASYN ensemble attained 97.5% accuracy and recall, with 97.4% precision and F1-score. AMSVM recorded 86.8% accuracy, 74.7% precision, 83.6% recall, and a 79% F1-score. This study improves IWQI classification, explores trade-offs between accuracy and class balance, and provides information on the effectiveness of class weighting, apportioned margins, and resampling techniques for the development of the ML model. The proposed models can facilitate the development of a low-cost IoT-based IWQI assessment system, supporting sustainable irrigation management to enhance agricultural productivity.

An adaptive synthetic sampling and batch generation-oriented hybrid approach for addressing class imbalance problem in software defect prediction

An adaptive synthetic sampling and batch generation-oriented hybrid approach for addressing class imbalance problem in software defect prediction

Accurate modeling of crude oil and brine interfacial tension via robust machine learning approaches

Interfacial tension (IFT) between water and crude oil is a crucial variable that enhanced oil recovery (EOR) techniques can adjust to increase oil extraction from depleted fields. Most of the developed intelligent models in the literature are based on synthetic oil samples rather than real crude oil samples or brine total salinity rather than salinity of each salt type. Hence, this study applies various machine learning approaches, such as Convolutional Neural Networks (CNN), Adaptive Boosting (AdaBoost), Decision Trees (DT), Random Forest (RF), K-Nearest Neighbors (KNN), Ensemble Learning, Support Vector Machines (SVM), and Multi-Layer Perceptron Artificial Neural Networks (MLP-ANN) to develop advanced models for predicting the IFT between brine and crude oil considering real crude oil samples and taking the account of each salt type prevalent within the brine phase, which represent the realistic circumstances encountered in the oil reservoirs. These predictions are based on factors like the type and concentration of salt, the API of the crude oil, and the properties of the system (pressure and temperature) using previously published experimental data. A sensitivity analysis, incorporating a relevancy factor, is performed to highlight the influence of various input parameters on the IFT. Among these models, the Decision Tree is highlighted for its high accuracy and low training cost compared to ANN-based models, as evidenced by its emerged evaluation metrics (R-squared of 0.9796 and mean square error of 5e-4). It is noted that the AdaBoost model is the least accurate with an R2 of 0.6696. Furthermore, the sensitivity analysis indicates that the molecular weight of the salt has the smallest impact on the IFT, whereas temperature has the most significant effect. The developed smart model may be used to accurately estimate crude oil/brine IFT without needing tedious, time-consuming and expensive experimental workflows.

Synthesis of calcium phosphate nanomaterial from quail eggshell for cadmium removal from wastewater and its genotoxic/cytotoxic properties.

In recent years, pollutants released into the environment from various sources threaten environmental health. Rapid industrialization and the constantly increasing needs of people facilitate the release of more hazardous wastes into the ecosystem. The presence of pollutants in water resources causes a wide range of adverse effects. In this study, calcium phosphate nanomaterial (Ca3(PO4)2 NM) was synthesized from biological waste eggshells for the cadmium removal in synthetic domestic wastewater, and a treatment method was developed using these NMs. The Ca3(PO4)2 NMs were produced by using a biowaste which provides the synthesis procedure greener approach. The biogenic NMs were used to remove toxic cadmium ions from wastewater samples. Cytotoxicity and genotoxicity studies of the synthesized NMs were also carried out, and their possible effects on the health of living organisms and the ecology were examined. In the developed method, the parameters affecting the removal of cadmium from wastewater samples were optimized and the removal efficiency was calculated by determining cadmium in a flame atomic absorption spectrophotometer system (FAAS). Synthetic domestic wastewater samples were utilized for evaluating the applicability of the developed treatment strategy. In addition, the adsorption capacity of the material for Cd2+ ion was calculated and the values obtained were modeled by using Langmuir adsorption isotherm (LAI). The calculated LAI parameters were within the appropriate limits, which proved that the developed NM can be used as an effective material for cadmium removal. Moreover, a new, rapid, and feasible synthesis strategy for the synthesis of Ca3(PO4)2 NM was presented in the literature.

Advanced Predictive Modeling for Hepatitis C Diagnosis Using Machine Learning

The hepatitis C virus (HCV) virus, which has infected nearly fifty million individuals across the world, has recently been receiving significant attention. With the advent of artificial intelligence, there is an increasing effort to predict the HCV by utilizing machine learning models and training such models on relevant datasets. These data-driven models can aid doctors in the early detection of Hepatitis C infection. In this study, using datasets from the National Center for Health Statistics, we aim to develop a reliable model for predicting patients with positive hepatitis C tests. The most significant challenge in predicting such diseases using machine learning methods is finding an appropriate technique to address extremely imbalanced datasets, which arise from the smaller number of patients compared to healthy individuals in real-world datasets. We utilized real-world datasets and employed different machine learning algorithms, along with various methods for data balancing and feature selection. Unlike previous studies, which lacked this level of comprehensiveness, we utilized diverse feature selection and sampling methods such as Recursive Feature Elimination, Analysis of Variance (ANOVA) feature selection, Correlation matrix, synthetic minority over-sampling technique (SMOTE), BorderlineSMOTE, support vector machine synthetic minority over-sampling technique (SVMSMOTE), and Adaptive Synthetic Sampling (ADASYN) in conjunction with different machine learning algorithms including Random Forest, Decision Tree, XGBoost, AdaBoost, and logistic regression. We are the first to develop such a comprehensive study for the prediction of the HCV. Our findings suggest that we can predict patients infected by the HCV with a reasonable recall score of 0.86. This recall is achieved through the application of a model using the AdaBoost algorithm and the ADASYN method for balancing the dataset. The model's performance, especially on the validation dataset, was slightly better when we used ANOVA feature selection. Feature importance analysis conducted with ANOVA feature selection and Recursive Feature Elimination indicates that four to five features are the most prominent in our research analysis to predict patients with positive Hepatitis C tests.

An Intelligent Kick Detection Model for Large-Hole Ultra-Deep Wells in the Sichuan Basin

The Sichuan Basin has abundant deep and ultra-deep natural gas resources, making it a primary target for exploration and the development of China’s oil and gas industry. However, during the drilling of ultra-deep wells in the Sichuan Basin, complex geological conditions frequently lead to gas kicks, posing significant challenges to well control and safety. Compared to traditional kick detection methods, artificial intelligence technology can improve the accuracy and timeliness of kick detection. However, there are limited real-world kick data available from drilling operations, and the datasets are extremely imbalanced, making it difficult to train intelligent models with sufficient accuracy and generalization capabilities. To address this issue, this paper proposes a kick data augmentation method based on a time-series generative adversarial network (TimeGAN). This method generates synthetic kick samples from real datasets and then employs a long short-term memory (LSTM) neural network to extract multivariate time-series features of surface drilling parameters. A multilayer perceptron (MLP) network is used for data classification tasks, constructing an intelligent kick detection model. Using real drilling data from ultra-deep wells in the SY block of the Sichuan Basin, the effects of k-fold cross-validation, data dimensionality, various imbalanced data handling techniques, and the sample imbalance ratio on the model’s kick detection performance are analyzed. Ablation experiments are also conducted to assess the contribution of each module in identifying kick. The results show that TimeGAN outperforms other imbalanced data handling techniques. The accuracy, recall, precision, and F1-score of the kick identification model are highest when the sample imbalance ratio is at 1 but decrease as the imbalance ratio increases. This indicates that maintaining a balance between positive and negative samples is essential for training a reliable intelligent kick detection model. The trained model is applied during the drilling of seven ultra-deep wells in Sichuan, demonstrating its effectiveness and accuracy in real-world kick detection.