Machine learning for the adsorptive removal of ciprofloxacin using sugarcane bagasse as a low-cost biosorbent: comparison of analytic, mechanistic, and neural network modeling.

Contamination with traces of pharmaceutical compounds, such as ciprofloxacin, has prompted interest in their removal via low-cost, efficient biomass-based adsorption. In this study, classical models, a mechanistic model, and a neural network model were evaluated for predicting ciprofloxacin breakthrough curves in both laboratory- and pilot scales. For the laboratory-scale (d = 2.2 cm, Co = 5 mg/L, Q = 7 mL/min, T = 18 °C) and pilot-scale (D = 4.4 cm, Co = 5 mg/L, Q = 28 mL/min, T = 18 °C) setups, the experimental adsorption capacities were 2.19 and 2.53 mg/g, respectively. The mechanistic model reproduced the breakthrough data with high accuracy on both scales (R2 > 0.4 and X2 < 0.15), and its fit was higher than conventional analytical models, namely the Clark, Modified Dose-Response, and Bohart-Adams models. The neural network model showed the highest level of agreement between predicted and experimental data with values of R2 = 0.993, X2 = 0.0032 (pilot-scale) and R2 = 0.986, X2 = 0.0022 (laboratory-scale). This study demonstrates that machine learning algorithms exhibit great potential for predicting the liquid adsorption of emerging pollutants in fixed bed.

Particulate matter forecast and prediction in Curitiba using machine learning.

Introduction: Air quality is directly affected by pollutant emission from vehicles, especially in large cities and metropolitan areas or when there is no compliance check for vehicle emission standards. Particulate Matter (PM) is one of the pollutants emitted from fuel burning in internal combustion engines and remains suspended in the atmosphere, causing respiratory and cardiovascular health problems to the population. In this study, we analyzed the interaction between vehicular emissions, meteorological variables, and particulate matter concentrations in the lower atmosphere, presenting methods for predicting and forecasting PM2.5. Methods: Meteorological and vehicle flow data from the city of Curitiba, Brazil, and particulate matter concentration data from optical sensors installed in the city between 2020 and 2022 were organized in hourly and daily averages. Prediction and forecasting were based on two machine learning models: Random Forest (RF) and Long Short-Term Memory (LSTM) neural network. The baseline model for prediction was chosen as the Multiple Linear Regression (MLR) model, and for forecast, we used the naive estimation as baseline. Results: RF showed that on hourly and daily prediction scales, the planetary boundary layer height was the most important variable, followed by wind gust and wind velocity in hourly or daily cases, respectively. The highest PM prediction accuracy (99.37%) was found using the RF model on a daily scale. For forecasting, the highest accuracy was 99.71% using the LSTM model for 1-h forecast horizon with 5 h of previous data used as input variables. Discussion: The RF and LSTM models were able to improve prediction and forecasting compared with MLR and Naive, respectively. The LSTM was trained with data corresponding to the period of the COVID-19 pandemic (2020 and 2021) and was able to forecast the concentration of PM2.5 in 2022, in which the data show that there was greater circulation of vehicles and higher peaks in the concentration of PM2.5. Our results can help the physical understanding of factors influencing pollutant dispersion from vehicle emissions at the lower atmosphere in urban environment. This study supports the formulation of new government policies to mitigate the impact of vehicle emissions in large cities.

Non-Invasive Prediction of Choledocholithiasis Using 1D Convolutional Neural Networks and Clinical Data.

This paper introduces a novel one-dimensional convolutional neural network that utilizes clinical data to accurately detect choledocholithiasis, where gallstones obstruct the common bile duct. Swift and precise detection of this condition is critical to preventing severe complications, such as biliary colic, jaundice, and pancreatitis. This cutting-edge model was rigorously compared with other machine learning methods commonly used in similar problems, such as logistic regression, linear discriminant analysis, and a state-of-the-art random forest, using a dataset derived from endoscopic retrograde cholangiopancreatography scans performed at Olive View-University of California, Los Angeles Medical Center. The one-dimensional convolutional neural network model demonstrated exceptional performance, achieving 90.77% accuracy and 92.86% specificity, with an area under the curve of 0.9270. While the paper acknowledges potential areas for improvement, it emphasizes the effectiveness of the one-dimensional convolutional neural network architecture. The results suggest that this one-dimensional convolutional neural network approach could serve as a plausible alternative to endoscopic retrograde cholangiopancreatography, considering its disadvantages, such as the need for specialized equipment and skilled personnel and the risk of postoperative complications. The potential of the one-dimensional convolutional neural network model to significantly advance the clinical diagnosis of this gallstone-related condition is notable, offering a less invasive, potentially safer, and more accessible alternative.

Multiclass Classification of Visual Electroencephalogram Based on Channel Selection, Minimum Norm Estimation Algorithm, and Deep Network Architectures.

This work addresses the challenge of classifying multiclass visual EEG signals into 40 classes for brain-computer interface applications using deep learning architectures. The visual multiclass classification approach offers BCI applications a significant advantage since it allows the supervision of more than one BCI interaction, considering that each class label supervises a BCI task. However, because of the nonlinearity and nonstationarity of EEG signals, using multiclass classification based on EEG features remains a significant challenge for BCI systems. In the present work, mutual information-based discriminant channel selection and minimum-norm estimate algorithms were implemented to select discriminant channels and enhance the EEG data. Hence, deep EEGNet and convolutional recurrent neural networks were separately implemented to classify the EEG data for image visualization into 40 labels. Using the k-fold cross-validation approach, average classification accuracies of 94.8% and 89.8% were obtained by implementing the aforementioned network architectures. The satisfactory results obtained with this method offer a new implementation opportunity for multitask embedded BCI applications utilizing a reduced number of both channels (<50%) and network parameters (<110 K).

High-gamma and beta bursts in the Left Supramarginal Gyrus can accurately differentiate verbal memory states and performance.

The left supramarginal gyrus (LSMG) may mediate attention to memory, and gauge memory state and performance. We performed a secondary analysis of 142 verbal delayed free recall experiments, in patients with medically-refractory epilepsy with electrode contacts implanted in the LSMG. In 14 of 142 experiments (in 14 of 113 patients), the cross-validated convolutional neural networks (CNNs) that used 1-dimensional(1-D) pairs of convolved high-gamma and beta tensors, derived from the LSMG recordings, could label recalled words with an area under the receiver operating curve (AUROC) of greater than 60% [range: 60-90%]. These 14 patients were distinguished by: 1) higher amplitudes of high-gamma bursts; 2) distinct electrode placement within the LSMG; and 3) superior performance compared with a CNN that used a 1-D tensor of the broadband recordings in the LSMG. In a pilot study of 7 of these patients, we also cross-validated CNNs using paired 1-D convolved high-gamma and beta tensors, from the LSMG, to: a) distinguish word encoding epochs from free recall epochs [AUC 0.6-1]; and distinguish better performance from poor performance during delayed free recall [AUC 0.5-0.86]. These experiments show that bursts of high-gamma and beta generated in the LSMG are biomarkers of verbal memory state and performance.

The role of artificial intelligence and data science in nanoparticles development: a review.

Artificial intelligence has revolutionized many sectors with unparalleled predictive capabilities supported by machine learning (ML). So far, this tool has not been able to provide the same level of development in pharmaceutical nanotechnology. This review discusses the current data science methodologies related to polymeric drug-loaded nanoparticle production from an innovative multidisciplinary perspective while considering the strictest data science practices. Several methodological and data interpretation flaws were identified by analyzing the few qualified ML studies. Most issues lie in following appropriate analysis steps, such as cross-validation, balancing data, or testing alternative models. Thus, better-planned studies following the recommended data science analysis steps along with adequate numbers of experiments would change the current landscape, allowing the exploration of the full potential of ML.

[Box: see text].

ANN-Based LiDAR Positioning System for B5G.

This work reports the development of an efficient and precise indoor positioning system utilizing two-dimensional (2D) light detection and ranging (LiDAR) technology, aiming to address the challenging sensing and positioning requirements of the beyond fifth-generation (B5G) mobile networks. The core of this work is the implementation of a 2D-LiDAR system enhanced by an artificial neural network (ANN), chosen due to its robustness against electromagnetic interference and higher accuracy over traditional radiofrequency signal-based methods. The proposed system uses 2D-LiDAR sensors for data acquisition and digital filters for signal improvement. Moreover, a camera and an image-processing algorithm are used to automate the labeling of samples that will be used to train the ANN by means of indicating the regions where the pedestrians are positioned. This accurate positioning information is essential for the optimization of B5G network operation, including the control of antenna arrays and reconfigurable intelligent surfaces (RIS). The experimental validation demonstrates the efficiency of mapping pedestrian locations with a precision of up to 98.787%, accuracy of 95.25%, recall of 98.537%, and an F1 score of 98.571%. These results show that the proposed system has the potential to solve the problem of sensing and positioning in indoor environments with high reliability and accuracy.

Modeling and optimization of microwave-assisted extraction of total phenolics content from mango (Mangifera indica) peel using response surface methodology (RSM) and artificial neural networks (ANN).

Mango (Mangifera indica) is a fruit highly consumed for its flavor and nutrient content. The mango peel is rich in compounds with biological functionality, such as antioxidant activity among others. The influence of microwave-assisted extraction variables on total phenol compounds (TPC) and antioxidant activity (TEAC) of natural extracts obtained from mango peel var. Tommy and Sugar were studied using a response surface methodology (RSM) and Artificial Neural Networks (ANN). TPC of mango peel extract var. Tommy was significantly influenced by time extraction (X1), solvent/plant ratio (X2) and concentration of ethanol (X3) and while mango peel extract var. Sugar was influenced by X2. TEAC by ABTS was significantly influenced by X3. Maximum of TPC (121.3 mg GAE / g of extract) and TEAC (1185.9 µmol Trolox/g extract) for mango peel extract var. Tommy were obtained at X1=23.9s, X2=12.6mL/gand X3=63.2%, and for mango peel extract var. Sugar, the maximum content of TPC (224.86 mg GAE/g extract) and TEAC (2117.7 µmol Trolox/g extract) were obtained at X1=40s, X2=10mL/g and X3=74.9%. The ANN model presented a higher predictive capacity than the RSM (RANN2>RRSM2,RMSEANN

A systematic evaluation of Euclidean alignment with deep learning for EEG decoding.

Objective:Electroencephalography signals are frequently used for various Brain-Computer interface (BCI) tasks. While deep learning (DL) techniques have shown promising results, they are hindered by the substantial data requirements. By leveraging data from multiple subjects, transfer learning enables more effective training of DL models. A technique that is gaining popularity is Euclidean alignment (EA) due to its ease of use, low computational complexity, and compatibility with DL models. However, few studies evaluate its impact on the training performance of shared and individual DL models. In this work, we systematically evaluate the effect of EA combined with DL for decoding BCI signals.Approach:We used EA as a pre-processing step to train shared DL models with data from multiple subjects and evaluated their transferability to new subjects.Main results:Our experimental results show that it improves decoding in the target subject by 4.33% and decreases convergence time by more than 70%. We also trained individual models for each subject to use as a majority-voting ensemble classifier. In this scenario, using EA improved the 3-model ensemble accuracy by 3.71%. However, when compared to the shared model with EA, the ensemble accuracy was 3.62% lower.Significance:EA succeeds in the task of improving transfer learning performance with DL models and, could be used as a standard pre-processing technique.

Artificial intelligence and radiomics in the diagnosis of intraosseous lesions of the gnathic bones: A systematic review.

BACKGROUND: The purpose of this systematic review (SR) is to gather evidence on the use of machine learning (ML) models in the diagnosis of intraosseous lesions in gnathic bones and to analyze the reliability, impact, and usefulness of such models. This SR was performed in accordance with the PRISMA 2022 guidelines and was registered in the PROSPERO database (CRD42022379298). METHODS: The acronym PICOS was used to structure the inquiry-focused review question "Is Artificial Intelligence reliable for the diagnosis of intraosseous lesions in gnathic bones?" The literature search was conducted in various electronic databases, including PubMed, Embase, Scopus, Cochrane Library, Web of Science, Lilacs, IEEE Xplore, and Gray Literature (Google Scholar and ProQuest). Risk of bias assessment was performed using PROBAST, and the results were synthesized by considering the task and sampling strategy of the dataset. RESULTS: Twenty-six studies were included (21 146 radiographic images). Ameloblastomas, odontogenic keratocysts, dentigerous cysts, and periapical cysts were the most frequently investigated lesions. According to TRIPOD, most studies were classified as type 2 (randomly divided). The F1 score was presented in only 13 studies, which provided the metrics for 20 trials, with a mean of 0.71 (±0.25). CONCLUSION: There is no conclusive evidence to support the usefulness of ML-based models in the detection, segmentation, and classification of intraosseous lesions in gnathic bones for routine clinical application. The lack of detail about data sampling, the lack of a comprehensive set of metrics for training and validation, and the absence of external testing limit experiments and hinder proper evaluation of model performance.

Integrated AI-driven optimization of Fenton process for the treatment of antibiotic sulfamethoxazole: Insights into mechanistic approach.

Antibiotics, as a class of environmental pollutants, pose a significant challenge due to their persistent nature and resistance to easy degradation. This study delves into modeling and optimizing conventional Fenton degradation of antibiotic sulfamethoxazole (SMX) and total organic carbon (TOC) under varying levels of H2O2, Fe2+ concentration, pH, and temperature using statistical and artificial intelligence techniques including Multiple Regression Analysis (MRA), Support Vector Regression (SVR) and Artificial Neural Network (ANN). In statistical metrics, the ANN model demonstrated superior predictive accuracy compared to its counterparts, with lowest RMSE values of 0.986 and 1.173 for SMX and TOC removal, respectively. Sensitivity showcased H2O2/Fe2+ ratio, time and pH as pivotal for SMX degradation, while in simultaneous SMX and TOC reduction, fine tuning the time, pH, and temperature was essential. Leveraging a Hybrid Genetic Algorithm-Desirability Optimization approach, the trained ANN model revealed an optimal desirability of 0.941 out of 1000 solutions which yielded a 91.18% SMX degradation and 87.90% TOC removal under following specific conditions: treatment time of 48.5 min, Fe2+: 7.05 mg L-1, H2O2: 128.82 mg L-1, pH: 5.1, initial SMX: 97.6 mg L-1, and a temperature: 29.8 °C. LC/MS analysis reveals multiple intermediates with higher m/z (242, 270 and 288) and lower m/z (98, 108, 156 and 173) values identified, however no aliphatic hydrocarbon was isolated, because of the low mineralization performance of Fenton process. Furthermore, some inorganic fragments like NH4+ and NO3- were also determined in solution. This comprehensive research enriches AI modeling for intricate Fenton-based contaminant degradation, advancing sustainable antibiotic removal strategies.

The effect of hair removal and filtering on melanoma detection: a comparative deep learning study with AlexNet CNN.

Melanoma is the most aggressive and prevalent form of skin cancer globally, with a higher incidence in men and individuals with fair skin. Early detection of melanoma is essential for the successful treatment and prevention of metastasis. In this context, deep learning methods, distinguished by their ability to perform automated and detailed analysis, extracting melanoma-specific features, have emerged. These approaches excel in performing large-scale analysis, optimizing time, and providing accurate diagnoses, contributing to timely treatments compared to conventional diagnostic methods. The present study offers a methodology to assess the effectiveness of an AlexNet-based convolutional neural network (CNN) in identifying early-stage melanomas. The model is trained on a balanced dataset of 10,605 dermoscopic images, and on modified datasets where hair, a potential obstructive factor, was detected and removed allowing for an assessment of how hair removal affects the model's overall performance. To perform hair removal, we propose a morphological algorithm combined with different filtering techniques for comparison: Fourier, Wavelet, average blur, and low-pass filters. The model is evaluated through 10-fold cross-validation and the metrics of accuracy, recall, precision, and the F1 score. The results demonstrate that the proposed model performs the best for the dataset where we implemented both a Wavelet filter and hair removal algorithm. It has an accuracy of 91.30%, a recall of 87%, a precision of 95.19%, and an F1 score of 90.91%.

Predicting intensive care unit-acquired weakness: A multilayer perceptron neural network approach.

In this editorial, we comment on the article by Wang and Long, published in a recent issue of the World Journal of Clinical Cases. The article addresses the challenge of predicting intensive care unit-acquired weakness (ICUAW), a neuromuscular disorder affecting critically ill patients, by employing a novel processing strategy based on repeated machine learning. The editorial presents a dataset comprising clinical, demographic, and laboratory variables from intensive care unit (ICU) patients and employs a multilayer perceptron neural network model to predict ICUAW. The authors also performed a feature importance analysis to identify the most relevant risk factors for ICUAW. This editorial contributes to the growing body of literature on predictive modeling in critical care, offering insights into the potential of machine learning approaches to improve patient outcomes and guide clinical decision-making in the ICU setting.

Morphological heterogeneity of neurons in the human central amygdaloid nucleus.

The central amygdaloid nucleus (CeA) has an ancient phylogenetic development and functions relevant for animal survival. Local cells receive intrinsic amygdaloidal information that codes emotional stimuli of fear, integrate them, and send cortical and subcortical output projections that prompt rapid visceral and social behavior responses. We aimed to describe the morphology of the neurons that compose the human CeA (N = 8 adult men). Cells within CeA coronal borders were identified using the thionine staining and were further analyzed using the "single-section" Golgi method followed by open-source software procedures for two-dimensional and three-dimensional image reconstructions. Our results evidenced varied neuronal cell body features, number and thickness of primary shafts, dendritic branching patterns, and density and shape of dendritic spines. Based on these criteria, we propose the existence of 12 morphologically different spiny neurons in the human CeA and discuss the variability in the dendritic architecture within cellular types, including likely interneurons. Some dendritic shafts were long and straight, displayed few collaterals, and had planar radiation within the coronal neuropil volume. Most of the sampled neurons showed a few to moderate density of small stubby/wide spines. Long spines (thin and mushroom) were observed occasionally. These novel data address the synaptic processing and plasticity in the human CeA. Our morphological description can be combined with further transcriptomic, immunohistochemical, and electrophysiological/connectional approaches. It serves also to investigate how neurons are altered in neurological and psychiatric disorders with hindered emotional perception, in anxiety, following atrophy in schizophrenia, and along different stages of Alzheimer's disease.

CNN application for automated determination of the patient's size to obtain the size-specific dose estimated in CT.

Introduction. The currently available dosimetry techniques in computed tomography can be inaccurate which overestimate the absorbed dose. Therefore, we aimed to provide an automated and fast methodology to more accurately calculate the SSDE usingDwobtained by using CNN from thorax and abdominal CT study images.Methods. The SSDE was determined from the 200 records files. For that purpose, patients' size was measured in two ways: (a) by developing an algorithm following the AAPM Report No. 204 methodology; and (b) using a CNN according to AAPM Report No. 220.Results. The patient's size measured by the in-house software in the region of thorax and abdomen was 27.63 ± 3.23 cm and 28.66 ± 3.37 cm, while CNN was 18.90 ± 2.6 cm and 21.77 ± 2.45 cm. The SSDE in thorax according to 204 and 220 reports were 17.26 ± 2.81 mGy and 23.70 ± 2.96 mGy for women and 17.08 ± 2.09 mGy and 23.47 ± 2.34 mGy for men. In abdomen was 18.54 ± 2.25 mGy and 23.40 ± 1.88 mGy in women and 18.37 ± 2.31 mGy and 23.84 ± 2.36 mGy in men.Conclusions. Implementing CNN-based automated methodologies can contribute to fast and accurate dose calculations, thereby improving patient-specific radiation safety in clinical practice.

Built environment and physical activity of adolescents: an approach with artificial neural networks / Ambiente construído e atividade física de adolescentes: uma abordagem com redes neurais artificiais

The objective of this study was to analyze the association of the level of physical activity (PA) and body composition in relation to the amount and distance of built environments favorable to the practice of PA in relation to the homes of adolescents in the city of Lagoa do Carro/Pernambuco, Brazil. A total of 289 adolescents (153 boys; 10 to 18 years) participated in the study, duly enrolled in schools municipality. The self-administered Physical Activity Questionnaire for Children (PAQ-C) and Physical Activity Questionnaire for Adolescent (PAQ-A) was used to assess the PA level. The Geographic Information System was used to assess the built environments. Buffers of 100 to 500 meters were created from the center of the built environment. The Artificial Neural Network in the Feedforward method was used to assess the association and importance of built environment and body composition variables with PA level. The different distances from the built environment to the place of residence do not present statistical differences. It is noteworthy that the number of buffers up to 500 meters away was the variable that showed the greatest importance for the PA level, along with adolescents who live in places with greater exposure in terms of built environments, being considered more active. It was possible to conclude that the main determinants of the PA level of adolescents were the amount of buffers at 500 meters, sex and the distance to the built environment. However, the variables, housing area, body mass and amounts of buffers at 100 meters were the ones with the lowest power of influence.

O objetivo deste estudo foi analisar a associação do nível de atividade física (AF) e composição corporal em relação à quantidade e distância de ambientes construídos favoráveis à prática da AF em relação ao domicílio de adolescentes da cidade de Lagoa do Carro/Pernambuco, Brasil. Participaram do estudo 289 adolescentes (153 meninos; 10 a 18 anos), devidamente matriculados nas escolas do município. O Physical Activity Questionnaire for Children (PAQ-C) e Physical Activity Questionnaire for Adolescent (PAQ-A) autoaplicável foram utilizados para avaliar o nível de AF. O Sistema de Informação Geográfico foi utilizado para avaliação dos ambientes construídos. Foram criados Buffers de 100 a 500 metros de raio a partir do centro do ambiente construído. A Rede Neural Artificial no método de Feedforward foi utilizada para analisar a associação e a importância das variáveis do ambiente construído e composição corporal com o nível de AF. Não foram observadas diferenças estatisticamente significativas entre o nível de AF e as distâncias do ambiente construído. Ressalta--se que a quantidade de buffers até 500 metros de distância, foi a variável que apresentou maior importância para o nível de AF, juntamente com os adolescentes que residem em locais com maior exposição em quantidade de ambientes construídos, sendo considerados mais ativos. Os principais determinantes do nível da AF dos adolescentes foram à quantidade de buffers a 500 metros, o sexo e a distância para o ambiente construído. Em contrapartida, as variáveis, zona de moradia, massa corporal e quantidades de buffers a 100 metros foram as que apresentaram um menor poder de influência.

Antioxidant Responses and Phytochemical Accumulation in Raphanus Species Sprouts through Elicitors and Predictive Models under High Temperature Stress.

Crop production is being impacted by higher temperatures, which can decrease food yield and pose a threat to human nutrition. In the current study, edible and wild radish sprouts were exposed to elevated growth temperatures along with the exogenous application of various elicitors to activate defense mechanisms. Developmental traits, oxidative damage, glucosinolate and anthocyanin content, and antioxidant capacity were evaluated alongside the development of a predictive model. A combination of four elicitors (citric acid, methyl jasmonate-MeJa, chitosan, and K2SO4) and high temperatures were applied. The accumulation of bioactives was significantly enhanced through the application of two elicitors, K2SO4 and methyl jasmonate (MeJa). The combination of high temperature with MeJa prominently activated oxidative mechanisms. Consequently, an artificial neural network was developed to predict the behavior of MeJa and temperature, providing a valuable projection of plant growth responses. This study demonstrates that the use of elicitors and predictive analytics serves as an effective tool to investigate responses and enhance the nutritional value of Raphanus species sprouts under future conditions of increased temperature.

Application of Microfracture Analysis to Fatigue Fractures in Materials through Non-Destructive Tests.

Fatigue fractures in materials are the main cause of approximately 80% of all material failures, and it is believed that such failures can be predicted and mathematically calculated in a reliable manner. It is possible to establish prediction modalities in cases of fatigue fractures according to three fundamental variables in fatigue, such as volume, number of fracture cycles, as well as applied stress, with the integration of Weibull constants (length characteristic). In this investigation, mechanical fatigue tests were carried out on specimens smaller than 4 mm2, made of different industrial materials. Their subsequent analysis was performed through precision computed tomography, in search for microfractures. The measurement of these microfractures, along with their metrics and classifications, was recorded. A convolutional neural network trained with deep learning was used to achieve the detection of microfractures in image processing. The detection of microfractures in images with resolutions of 480 × 854 or 960 × 960 pixels is the primary objective of this network, and its accuracy is above 95%. Images that have microfractures and those without are classified using the network. Subsequently, by means of image processing, the microfracture is isolated. Finally, the images containing this feature are interpreted using image processing to obtain their area, perimeter, characteristic length, circularity, orientation, and microfracture-type metrics. All values are obtained in pixels and converted to metric units (µm) through a conversion factor based on image resolution. The growth of microfractures will be used to define trends in the development of fatigue fractures through the studies presented.

Convolutional Neural Networks in the Inspection of Serrasalmids (Characiformes) Fingerlings.

Aquaculture produces more than 122 million tons of fish globally. Among the several economically important species are the Serrasalmidae, which are valued for their nutritional and sensory characteristics. To meet the growing demand, there is a need for automation and accuracy of processes, at a lower cost. Convolutional neural networks (CNNs) are a viable alternative for automation, reducing human intervention, work time, errors, and production costs. Therefore, the objective of this work is to evaluate the efficacy of convolutional neural networks (CNNs) in counting round fish fingerlings (Serrasalmidae) at different densities using 390 color photographs in an illuminated environment. The photographs were submitted to two convolutional neural networks for object detection: one model was adapted from a pre-trained CNN and the other was an online platform based on AutoML. The metrics used for performance evaluation were precision (P), recall (R), accuracy (A), and F1-Score. In conclusion, convolutional neural networks (CNNs) are effective tools for detecting and counting fish. The pre-trained CNN demonstrated outstanding performance in identifying fish fingerlings, achieving accuracy, precision, and recall rates of 99% or higher, regardless of fish density. On the other hand, the AutoML exhibited reduced accuracy and recall rates as the number of fish increased.

An Interpretable Approach with Explainable AI for Heart Stroke Prediction.

Heart strokes are a significant global health concern, profoundly affecting the wellbeing of the population. Many research endeavors have focused on developing predictive models for heart strokes using ML and DL techniques. Nevertheless, prior studies have often failed to bridge the gap between complex ML models and their interpretability in clinical contexts, leaving healthcare professionals hesitant to embrace them for critical decision-making. This research introduces a meticulously designed, effective, and easily interpretable approach for heart stroke prediction, empowered by explainable AI techniques. Our contributions include a meticulously designed model, incorporating pivotal techniques such as resampling, data leakage prevention, feature selection, and emphasizing the model's comprehensibility for healthcare practitioners. This multifaceted approach holds the potential to significantly impact the field of healthcare by offering a reliable and understandable tool for heart stroke prediction. In our research, we harnessed the potential of the Stroke Prediction Dataset, a valuable resource containing 11 distinct attributes. Applying these techniques, including model interpretability measures such as permutation importance and explainability methods like LIME, has achieved impressive results. While permutation importance provides insights into feature importance globally, LIME complements this by offering local and instance-specific explanations. Together, they contribute to a comprehensive understanding of the Artificial Neural Network (ANN) model. The combination of these techniques not only aids in understanding the features that drive overall model performance but also helps in interpreting and validating individual predictions. The ANN model has achieved an outstanding accuracy rate of 95%.