Interpretation and explanation of computer vision classification of carambola (Averrhoa carambola L.) according to maturity stage.

The classification of carambola, also known as starfruit, according to quality parameters is usually conducted by trained human evaluators through visual inspections. This is a costly and subjective method that can generate high variability in results. As an alternative, computer vision systems (CVS) combined with deep learning (DCVS) techniques have been introduced in the industry as a powerful and an innovative tool for the rapid and non-invasive classification of fruits. However, validating the learning capability and trustworthiness of a DL model, aka black box, to obtain insights can be challenging. To reduce this gap, we propose an integrated eXplainable Artificial Intelligence (XAI) method for the classification of carambolas at different maturity stages. We compared two Residual Neural Networks (ResNet) and Visual Transformers (ViT) to identify the image regions that are enhanced by a Random Forest (RF) model, with the aim of providing more detailed information at the feature level for classifying the maturity stage. Changes in fruit colour and physicochemical data throughout the maturity stages were analysed, and the influence of these parameters on the maturity stages was evaluated using the Gradient-weighted Class Activation Mapping (Grad-CAM), the Attention Maps using RF importance. The proposed approach provides a visualization and description of the most important regions that led to the model decision, in wide visualization follows the models an importance features from RF. Our approach has promising potential for standardized and rapid carambolas classification, achieving 91 % accuracy with ResNet and 95 % with ViT, with potential application for other fruits.

Predicting the pulse of the Amazon: Machine learning insights into deforestation dynamics.

This study aims to analyze deforestation in the Brazilian Amazon from 1999 to 2020 using machine learning techniques to assess 16 critical factors. Our approach leverages the capabilities of machine learning, particularly Random Forest, which proved to be the most accurate model in terms of determination coefficient, mean squared error, and mean absolute error. The analysis revealed that the harvested area of permanent crops is the most influential variable in predicting deforestation, followed by the area of temporary crops. Furthermore, our findings indicate a significant inverse relationship between public spending and deforestation rates. These results contribute to understanding deforestation dynamics and offer potential strategies for improving conservation efforts.

Identifying Substructures That Facilitate Compounds to Penetrate the Blood-Brain Barrier via Passive Transport Using Machine Learning Explainer Models.

The local interpretable model-agnostic explanation (LIME) method was used to interpret two machine learning models of compounds penetrating the blood-brain barrier. The classification models, Random Forest, ExtraTrees, and Deep Residual Network, were trained and validated using the blood-brain barrier penetration dataset, which shows the penetrability of compounds in the blood-brain barrier. LIME was able to create explanations for such penetrability, highlighting the most important substructures of molecules that affect drug penetration in the barrier. The simple and intuitive outputs prove the applicability of this explainable model to interpreting the permeability of compounds across the blood-brain barrier in terms of molecular features. LIME explanations were filtered with a weight equal to or greater than 0.1 to obtain only the most relevant explanations. The results showed several structures that are important for blood-brain barrier penetration. In general, it was found that some compounds with nitrogenous substructures are more likely to permeate the blood-brain barrier. The application of these structural explanations may help the pharmaceutical industry and potential drug synthesis research groups to synthesize active molecules more rationally.

Evaluation of the Relation between Ictal EEG Features and XAI Explanations.

Epilepsy is a neurological disease with one of the highest rates of incidence worldwide. Although EEG is a crucial tool for its diagnosis, the manual detection of epileptic seizures is time consuming. Automated methods are needed to streamline this process; although there are already several works that have achieved this, the process by which it is executed remains a black box that prevents understanding of the ways in which machine learning algorithms make their decisions. A state-of-the-art deep learning model for seizure detection and three EEG databases were chosen for this study. The developed models were trained and evaluated under different conditions (i.e., three distinct levels of overlap among the chosen EEG data windows). The classifiers with the best performance were selected, then Shapley Additive Explanations (SHAPs) and Local Interpretable Model-Agnostic Explanations (LIMEs) were employed to estimate the importance value of each EEG channel and the Spearman's rank correlation coefficient was computed between the EEG features of epileptic signals and the importance values. The results show that the database and training conditions may affect a classifier's performance. The most significant accuracy rates were 0.84, 0.73, and 0.64 for the CHB-MIT, Siena, and TUSZ EEG datasets, respectively. In addition, most EEG features displayed negligible or low correlation with the importance values. Finally, it was concluded that a correlation between the EEG features and the importance values (generated by SHAP and LIME) may have been absent even for the high-performance models.

COVID-19 diagnosis using clinical markers and multiple explainable artificial intelligence approaches: A case study from Ecuador.

The COVID-19 pandemic erupted at the beginning of 2020 and proved fatal, causing many casualties worldwide. Immediate and precise screening of affected patients is critical for disease control. COVID-19 is often confused with various other respiratory disorders since the symptoms are similar. As of today, the reverse transcription-polymerase chain reaction (RT-PCR) test is utilized for diagnosing COVID-19. However, this approach is sometimes prone to producing erroneous and false negative results. Hence, finding a reliable diagnostic method that can validate the RT-PCR test results is crucial. Artificial intelligence (AI) and machine learning (ML) applications in COVID-19 diagnosis has proven to be beneficial. Hence, clinical markers have been utilized for COVID-19 diagnosis with the help of several classifiers in this study. Further, five different explainable artificial intelligence techniques have been utilized to interpret the predictions. Among all the algorithms, the k-nearest neighbor obtained the best performance with an accuracy, precision, recall and f1-score of 84%, 85%, 84% and 84%. According to this study, the combination of clinical markers such as eosinophils, lymphocytes, red blood cells and leukocytes was significant in differentiating COVID-19. The classifiers can be utilized synchronously with the standard RT-PCR procedure making diagnosis more reliable and efficient.

Dysbiosis signatures of gut microbiota and the progression of type 2 diabetes: a machine learning approach in a Mexican cohort.

Introduction: The gut microbiota (GM) dysbiosis is one of the causal factors for the progression of different chronic metabolic diseases, including type 2 diabetes mellitus (T2D). Understanding the basis that laid this association may lead to developing new therapeutic strategies for preventing and treating T2D, such as probiotics, prebiotics, and fecal microbiota transplants. It may also help identify potential early detection biomarkers and develop personalized interventions based on an individual's gut microbiota profile. Here, we explore how supervised Machine Learning (ML) methods help to distinguish taxa for individuals with prediabetes (prediabetes) or T2D. Methods: To this aim, we analyzed the GM profile (16s rRNA gene sequencing) in a cohort of 410 Mexican naïve patients stratified into normoglycemic, prediabetes, and T2D individuals. Then, we compared six different ML algorithms and found that Random Forest had the highest predictive performance in classifying T2D and prediabetes patients versus controls. Results: We identified a set of taxa for predicting patients with T2D compared to normoglycemic individuals, including Allisonella, Slackia, Ruminococus_2, Megaspgaera, Escherichia/Shigella, and Prevotella, among them. Besides, we concluded that Anaerostipes, Intestinibacter, Prevotella_9, Blautia, Granulicatella, and Veillonella were the relevant genus in patients with prediabetes compared to normoglycemic subjects. Discussion: These findings allow us to postulate that GM is a distinctive signature in prediabetes and T2D patients during the development and progression of the disease. Our study highlights the role of GM and opens a window toward the rational design of new preventive and personalized strategies against the control of this disease.

Geometric deep learning as a potential tool for antimicrobial peptide prediction.

Antimicrobial peptides (AMPs) are components of natural immunity against invading pathogens. They are polymers that fold into a variety of three-dimensional structures, enabling their function, with an underlying sequence that is best represented in a non-flat space. The structural data of AMPs exhibits non-Euclidean characteristics, which means that certain properties, e.g., differential manifolds, common system of coordinates, vector space structure, or translation-equivariance, along with basic operations like convolution, in non-Euclidean space are not distinctly established. Geometric deep learning (GDL) refers to a category of machine learning methods that utilize deep neural models to process and analyze data in non-Euclidean settings, such as graphs and manifolds. This emerging field seeks to expand the use of structured models to these domains. This review provides a detailed summary of the latest developments in designing and predicting AMPs utilizing GDL techniques and also discusses both current research gaps and future directions in the field.

Explaining and Visualizing Embeddings of One-Dimensional Convolutional Models in Human Activity Recognition Tasks.

Human Activity Recognition (HAR) is a complex problem in deep learning, and One-Dimensional Convolutional Neural Networks (1D CNNs) have emerged as a popular approach for addressing it. These networks efficiently learn features from data that can be utilized to classify human activities with high performance. However, understanding and explaining the features learned by these networks remains a challenge. This paper presents a novel eXplainable Artificial Intelligence (XAI) method for generating visual explanations of features learned by one-dimensional CNNs in its training process, utilizing t-Distributed Stochastic Neighbor Embedding (t-SNE). By applying this method, we provide insights into the decision-making process through visualizing the information obtained from the model's deepest layer before classification. Our results demonstrate that the learned features from one dataset can be applied to differentiate human activities in other datasets. Our trained networks achieved high performance on two public databases, with 0.98 accuracy on the SHO dataset and 0.93 accuracy on the HAPT dataset. The visualization method proposed in this work offers a powerful means to detect bias issues or explain incorrect predictions. This work introduces a new type of XAI application, enhancing the reliability and practicality of CNN models in real-world scenarios.

A Hybrid Algorithm of ML and XAI to Prevent Breast Cancer: A Strategy to Support Decision Making.

Worldwide, the coronavirus has intensified the management problems of health services, significantly harming patients. Some of the most affected processes have been cancer patients' prevention, diagnosis, and treatment. Breast cancer is the most affected, with more than 20 million cases and at least 10 million deaths by 2020. Various studies have been carried out to support the management of this disease globally. This paper presents a decision support strategy for health teams based on machine learning (ML) tools and explainability algorithms (XAI). The main methodological contributions are: first, the evaluation of different ML algorithms that allow classifying patients with and without cancer from the available dataset; and second, an ML methodology mixed with an XAI algorithm, which makes it possible to predict the disease and interpret the variables and how they affect the health of patients. The results show that first, the XGBoost Algorithm has a better predictive capacity, with an accuracy of 0.813 for the train data and 0.81 for the test data; and second, with the SHAP algorithm, it is possible to know the relevant variables and their level of significance in the prediction, and to quantify the impact on the clinical condition of the patients, which will allow health teams to offer early and personalized alerts for each patient.

Application of C5.0 Algorithm for the Assessment of Perceived Stress in Healthcare Professionals Attending COVID-19.

Coronavirus disease (COVID-19) represents one of the greatest challenges to public health in modern history. As the disease continues to spread globally, medical and allied healthcare professionals have become one of the most affected sectors. Stress and anxiety are indirect effects of the COVID-19 pandemic. Therefore, it is paramount to understand and categorize their perceived levels of stress, as it can be a detonating factor leading to mental illness. Here, we propose a computer-based method to better understand stress in healthcare workers facing COVID-19 at the beginning of the pandemic. We based our study on a representative sample of healthcare professionals attending to COVID-19 patients in the northeast region of Mexico, at the beginning of the pandemic. We used a machine learning classification algorithm to obtain a visualization model to analyze perceived stress. The C5.0 decision tree algorithm was used to study datasets. We carried out an initial preprocessing statistical analysis for a group of 101 participants. We performed chi-square tests for all questions, individually, in order to validate stress level calculation (p < 0.05) and a calculated Cronbach's alpha of 0.94 and McDonald's omega of 0.95, demonstrating good internal consistency in the dataset. The obtained model failed to classify only 6 out of the 101, missing two cases for mild, three for moderate and one for severe (accuracy of 94.1%). We performed statistical correlation analysis to ensure integrity of the method. In addition, based on the decision tree model, we concluded that severe stress cases can be related mostly to high levels of xenophobia and compulsive stress. Thus, showing that applied machine learning algorithms represent valuable tools in the assessment of perceived stress, which can potentially be adapted to other areas of the medical field.

CML-Cardio: a cascade machine learning model to predict cardiovascular disease risk as a primary prevention strategy.

Cardiovascular diseases are among the leading causes of mortality worldwide, with more than 23 million related deaths per year by 2030, according to the World Heart Federation. Although most of these diseases may be prevented, population awareness strategies are still ineffective. In this context, we propose the CML-Cardio tool, a machine learning application to automate the risk classification process of developing CVDs. For this, researchers in our group collected data on diabetes, blood pressure, and other risk factors in a private company. Our final model consists of a cascade system to handle highly imbalanced data. In the first stage, a binary model is responsible for predicting whether a patient has a low risk of developing CVDs or if has a risk that needs attention. In this step, we use six algorithms: logistic regression, SVM, random forest, XGBoost, CatBoost, and multilayer perceptron. The better results presented an average accuracy of 0.86 ± 0.03 and f-score of 0.85 ± 0.04. We interpret each feature's impact on the models' output and validate the subsystem for the next step. In the second stage, we use an anomaly detection model to learn the intermediate risk patterns present in the instances that need attention. The cascade model presented an average accuracy of 0.80 ± 0.07 and f-score of 0.70 ± 0.07. Finally, we develop the CML-Cardio prototype of an actual application as a primary prevention strategy. Graphical abstract In this work, we propose the CML-Cardio tool, a cascade machine learning method to classify cardiovascular disease risk.

Fine-scale modeling of the urban heat island: A comparison of multiple linear regression and random forest approaches.

Characterizing the spatiotemporal variability of the Urban Heat Island (UHI) and its drivers is a key step in leveraging thermal comfort to create not only healthier cities, but also to enhance urban resilience to climate change. In this study, we developed specific daytime and nighttime multiple linear regression (MLR) and random forest (RF) models to analyze and predict the spatiotemporal evolution of the Urban Heat Island intensity (UHII), using the air temperature (Tair) as the response variable. We profited from the wealth of in situ Tair data and a comprehensive pool of predictors variables - including land cover, population, traffic, urban geometry, weather data and atmospheric vertical indices. Cluster analysis divided the study period into three main groups, each dominated by a combination of weather systems that, in turn, influenced the onset and strength of the UHII. Anticyclonic circulations favored the emergence of the largest UHII (hourly mean of 5.06 °C), while cyclonic circulations dampened its development. The MLR models were only able to explain a modest percentage of variance (64 and 34% for daytime and nighttime, respectively), which we interpret as part of their inability to capture key factors controlling Tair. The RF models, on the other hand, performed considerably better, with explanatory power over 96% of the variance for daytime and nighttime conditions, capturing and mapping the fine-scale Tair spatiotemporal variability in both periods and under each cluster condition. The feature importance analysis showed that the meteorological variables and the land cover were the main predictors of the Tair. Urban planners could benefit from these results, using the high-performing RF models as a robust framework for forecasting and mitigating the effects of the UHI.

Impact of Lung Segmentation on the Diagnosis and Explanation of COVID-19 in Chest X-ray Images.

COVID-19 frequently provokes pneumonia, which can be diagnosed using imaging exams. Chest X-ray (CXR) is often useful because it is cheap, fast, widespread, and uses less radiation. Here, we demonstrate the impact of lung segmentation in COVID-19 identification using CXR images and evaluate which contents of the image influenced the most. Semantic segmentation was performed using a U-Net CNN architecture, and the classification using three CNN architectures (VGG, ResNet, and Inception). Explainable Artificial Intelligence techniques were employed to estimate the impact of segmentation. A three-classes database was composed: lung opacity (pneumonia), COVID-19, and normal. We assessed the impact of creating a CXR image database from different sources, and the COVID-19 generalization from one source to another. The segmentation achieved a Jaccard distance of 0.034 and a Dice coefficient of 0.982. The classification using segmented images achieved an F1-Score of 0.88 for the multi-class setup, and 0.83 for COVID-19 identification. In the cross-dataset scenario, we obtained an F1-Score of 0.74 and an area under the ROC curve of 0.9 for COVID-19 identification using segmented images. Experiments support the conclusion that even after segmentation, there is a strong bias introduced by underlying factors from different sources.

Brain Hemorrhage Classification in CT Scan Images Using Minimalist Machine Learning.

Over time, a myriad of applications have been generated for pattern classification algorithms. Several case studies include parametric classifiers such as the Multi-Layer Perceptron (MLP) classifier, which is one of the most widely used today. Others use non-parametric classifiers, Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), Naïve Bayes (NB), Adaboost, and Random Forest (RF). However, there is still little work directed toward a new trend in Artificial Intelligence (AI), which is known as eXplainable Artificial Intelligence (X-AI). This new trend seeks to make Machine Learning (ML) algorithms increasingly simple and easy to understand for users. Therefore, following this new wave of knowledge, in this work, the authors develop a new pattern classification methodology, based on the implementation of the novel Minimalist Machine Learning (MML) paradigm and a higher relevance attribute selection algorithm, which we call dMeans. We examine and compare the performance of this methodology with MLP, NB, KNN, SVM, Adaboost, and RF classifiers to perform the task of classification of Computed Tomography (CT) brain images. These grayscale images have an area of 128 × 128 pixels, and there are two classes available in the dataset: CT without Hemorrhage and CT with Intra-Ventricular Hemorrhage (IVH), which were classified using the Leave-One-Out Cross-Validation method. Most of the models tested by Leave-One-Out Cross-Validation performed between 50% and 75% accuracy, while sensitivity and sensitivity ranged between 58% and 86%. The experiments performed using our methodology matched the best classifier observed with 86.50% accuracy, and they outperformed all state-of-the-art algorithms in specificity with 91.60%. This performance is achieved hand in hand with simple and practical methods, which go hand in hand with this trend of generating easily explainable algorithms.

Explaining machine learning based diagnosis of COVID-19 from routine blood tests with decision trees and criteria graphs.

The sudden outbreak of coronavirus disease 2019 (COVID-19) revealed the need for fast and reliable automatic tools to help health teams. This paper aims to present understandable solutions based on Machine Learning (ML) techniques to deal with COVID-19 screening in routine blood tests. We tested different ML classifiers in a public dataset from the Hospital Albert Einstein, São Paulo, Brazil. After cleaning and pre-processing the data has 608 patients, of which 84 are positive for COVID-19 confirmed by RT-PCR. To understand the model decisions, we introduce (i) a local Decision Tree Explainer (DTX) for local explanation and (ii) a Criteria Graph to aggregate these explanations and portrait a global picture of the results. Random Forest (RF) classifier achieved the best results (accuracy 0.88, F1-score 0.76, sensitivity 0.66, specificity 0.91, and AUROC 0.86). By using DTX and Criteria Graph for cases confirmed by the RF, it was possible to find some patterns among the individuals able to aid the clinicians to understand the interconnection among the blood parameters either globally or on a case-by-case basis. The results are in accordance with the literature and the proposed methodology may be embedded in an electronic health record system.

Operational framework and training standard requirements for AI-empowered robotic surgery.

BACKGROUND: For autonomous robot-delivered surgeries to ever become a feasible option, we recommend the combination of human-centered artificial intelligence (AI) and transparent machine learning (ML), with integrated Gross anatomy models. This can be supplemented with medical imaging data of cadavers for performance evaluation. METHODS: We reviewed technological advances and state-of-the-art documented developments. We undertook a literature search on surgical robotics and skills, tracing agent studies, relevant frameworks, and standards for AI. This embraced transparency aspects of AI. CONCLUSION: We recommend "a procedure/skill template" for teaching AI that can be used by a surgeon. Similar existing methodologies show that when such a metric-based approach is used for training surgeons, cardiologists, and anesthetists, it results in a >40% error reduction in objectively assessed intraoperative procedures. The integration of Explainable AI and ML, and novel tissue characterization sensorics to tele-operated robotic-assisted procedures with medical imaged cadavers, provides robotic guidance and refines tissue classifications at a molecular level.