Search

The infection by SARS-CoV-2 which causes the COVID-19 disease has spread widely over the whole world since the beginning of 2020. Following the epidemic which started in Wuhan, China on January 30, 2020 the World Health Organization (WHO) declared a global health emergency and a pandemic. In this paper, we describe a publicly available multiclass CT scan dataset for SARS-CoV-2 infection identification. Which currently contains 4173 CT-scans of 210 different patients, out of which 2168 correspond to 80 patients infected with SARS-CoV-2 and confirmed by RT-PCR. These data have been collected in the Public Hospital of the Government Employees of Sao Paulo and the Metropolitan Hospital of Lapa, both in Sao Paulo; Brazil. The aim of this data set is to encourage the research and development of artificial intelligent methods that are able to identify SARS-CoV-2 or other diseases through the analysis of CT scans. As a baseline result for this data set, we used the recently introduced eXplainable Deep Learning approach (xDNN), which is a transparent deep learning approach that allows users to inspect the decisions of the network.

The infection by SARS-CoV-2 which causes the COVID-19 disease has spread widely over the whole world since the beginning of 2020. Following the epidemic which started in Wuhan, China on January 30, 2020 the World Health Organization (WHO) declared a global health emergency and a pandemic. In this paper, we describe a publicly available multiclass CT scan dataset for SARS-CoV-2 infection identification. Which currently contains 4173 CT-scans of 210 different patients, out of which 2168 correspond to 80 patients infected with SARS-CoV-2 and confirmed by RT-PCR. These data have been collected in the Public Hospital of the Government Employees of Sao Paulo and the Metropolitan Hospital of Lapa, both in Sao Paulo; Brazil. The aim of this data set is to encourage the research and development of artificial intelligent methods that are able to identify SARS-CoV-2 or other diseases through the analysis of CT scans. As a baseline result for this data set, we used the recently introduced eXplainable Deep Learning approach (xDNN), which is a transparent deep learning approach that allows users to inspect the decisions of the network.

Deep neural networks (DNN’s) have become essential for solving diverse complex problems and have achieved considerable success in tackling computer vision tasks. How-ever, DNN’s are vulnerable to human-imperceptible adversarial distortion/noise patterns that that can detrimentally impact safety-critical applications such as autonomous driving. In this paper, we introduce a novel robust-by-design deep learn-ing approach, Sim-DNN, that is able to detect adversarial attacks through its inner defense mechanism that considers the degree of similarity between new data samples and autonomously chosen prototypes. The approach benefits from the abrupt drop of the similarity score to detect concept changes caused by distorted/noise data when comparing their similarities against the set of prototypes. Due to the feed-forward prototype-based architecture of Sim-DNN, no re-training or adversarial training is required. In order to evaluate the robustness of the proposed method, we considered the recently introduced ImageNet-R dataset and different adversarial attack methods such as FGSM, PGD, and DDN. Different DNN’s methods were also considered in the analysis. Results have shown that the proposed Sim-DNN is able to detect adversarial attacks with better performance than its mainstream competitors. Moreover, as no adversarial training is required by Sim-DNN, its performance on clean and robust images is more stable than its competitors which require an external defense mechanism to improve their robustness.

Earth observation is fundamental for a range of human activities including flood response as it offers vital information to decision makers. Semantic segmentation plays a key role in mapping the raw hyper-spectral data coming from the satellites into a human understandable form assigning class labels to each pixel. Traditionally, water index based methods have been used for detecting water pixels. More recently, deep learning techniques such as U-Net started to gain attention offering significantly higher accuracy. However, the latter are hard to interpret by humans and use dozens of millions of abstract parameters that are not directly related to the physical nature of the problem being modelled. They are also labelled data and computational power hungry. At the same time, data transmission capability on small nanosatellites is limited in terms of power and bandwidth yet constellations of such small, nanosatellites are preferable, because they reduce the revisit time in disaster areas from days to hours. Therefore, being able to achieve as highly accurate models as deep learning (e.g. U-Net) or even more, to surpass them in terms of accuracy, but without the need to rely on huge amounts of labelled training data, computational power, abstract coefficients offers potentially game-changing capabilities for EO (Earth observation) and flood detection, in particular. In this paper, we introduce a prototype-based interpretable deep semantic segmentation (IDSS) method, which is highly accurate as well as interpretable. Its parameters are in orders of magnitude less than the number of parameters used by deep networks such as U-Net and are clearly interpretable by humans. The proposed here IDSS offers a transparent structure that allows users to inspect and audit the algorithm's decision. Results have demonstrated that IDSS could surpass other algorithms, including U-Net, in terms of IoU (Intersection over Union) total water and Recall total water. We used WorldFloods data

Deep neural networks (DNN’s) have become essential for solving diverse complex problems and have achieved considerable success in tackling computer vision tasks. How-ever, DNN’s are vulnerable to human-imperceptible adversarial distortion/noise patterns that that can detrimentally impact safety-critical applications such as autonomous driving. In this paper, we introduce a novel robust-by-design deep learn-ing approach, Sim-DNN, that is able to detect adversarial attacks through its inner defense mechanism that considers the degree of similarity between new data samples and autonomously chosen prototypes. The approach benefits from the abrupt drop of the similarity score to detect concept changes caused by distorted/noise data when comparing their similarities against the set of prototypes. Due to the feed-forward prototype-based architecture of Sim-DNN, no re-training or adversarial training is required. In order to evaluate the robustness of the proposed method, we considered the recently introduced ImageNet-R dataset and different adversarial attack methods such as FGSM, PGD, and DDN. Different DNN’s methods were also considered in the analysis. Results have shown that the proposed Sim-DNN is able to detect adversarial attacks with better performance than its mainstream competitors. Moreover, as no adversarial training is required by Sim-DNN, its performance on clean and robust images is more stable than its competitors which require an external defense mechanism to improve their robustness.

This paper describes a novel approach to the problem of developing explainable machine learning models. We consider a Deep Reinforcement Learning (DRL) model representing a highway path planning policy for autonomous highway driving. The model constitutes a mapping from the continuous multidimensional state space characterizing vehicle positions and velocities to a discrete set of actions in longitudinal and lateral direction. It is obtained by applying a customized version of the Double Deep Q-Network (DDQN) learning algorithm. The main idea is to approximate the DRL model with a set of IF…THEN rules that provide an alternative interpretable model, which is further enhanced by visualizing the rules. This concept is rationalized by the universal approximation properties of the rule-based models with fuzzy predicates. The proposed approach includes a learning engine composed of 0-order fuzzy rules, which generalize locally around the prototypes by using multivariate function models. The adjacent (in the data space) prototypes, which correspond to the same action are further grouped and merged into so-called "MegaClouds" reducing significantly the number of fuzzy rules. The input selection method is based on ranking the density of the individual inputs. Experimental results show that the specific DRL agent can be interpreted by approximating with families of rules of different granularity. The method is computationally efficient and can be potentially extended to addressing the explainability of the broader set of fully connected deep neural network models

This paper provides a brief analytical review of the current state-of-the-art in relation to the explainability of artificial intelligence in the context of recent advances in machine learning and deep learning. The paper starts with a brief historical introduction and a taxonomy, and formulates the main challenges in terms of explainability building on the recently formulated National Institute of Standards four principles of explainability. Recently published methods related to the topic are then critically reviewed and analyzed. Finally, future directions for research are suggested. This article is categorized under: Technologies > Artificial Intelligence Fundamental Concepts of Data and Knowledge > Explainable AI.

As COVID-19 hounds the world, the common cause of finding a swift solution to manage the pandemic has brought together researchers, institutions, governments, and society at large. The Internet of Things (IoT), Artificial Intelligence (AI) — including Machine Learning (ML) and Big Data analytics — as well as Robotics and Blockchain, are the four decisive areas of technological innovation that have been ingenuity harnessed to fight this pandemic and future ones. While these highly interrelated smart and connected health technologies cannot resolve the pandemic overnight and may not be the only answer to the crisis, they can provide greater insight into the disease and support frontline efforts to prevent and control the pandemic. This paper provides a blend of discussions on the contribution of these digital technologies, propose several complementary and multidisciplinary techniques to combat COVID-19, offer opportunities for more holistic studies, and accelerate knowledge acquisition and scientific discoveries in pandemic research. First, four areas where IoT can contribute are discussed, namely, i) tracking and tracing, ii) Remote Patient Monitoring (RPM) by Wearable IoT (WIoT), iii) Personal Digital Twins (PDT), and iv) real-life use case: ICT/IoT solution in Korea. Second, the role and novel applications of AI are explained, namely: i) diagnosis and prognosis, ii) risk prediction, iii) vaccine and drug development, iv) research dataset, v) early warnings and alerts, vi) social control and fake news detection, and vii) communication and chatbot. Third, the main uses of robotics and drone technology are analyzed, including i) crowd surveillance, ii) public announcements, iii) screening and diagnosis, and iv) essential supply delivery. Finally, we discuss how Distributed Ledger Technologies (DLTs), of which blockchain is a common example, can be combined with other technologies for tackling COVID-19.

This paper provides a brief analytical review of the current state-of-the-art in relation to the explainability of artificial intelligence in the context of recent advances in machine learning and deep learning. The paper starts with a brief historical introduction and a taxonomy, and formulates the main challenges in terms of explainability building on the recently formulated National Institute of Standards four principles of explainability. Recently published methods related to the topic are then critically reviewed and analyzed. Finally, future directions for research are suggested. This article is categorized under: Technologies > Artificial Intelligence Fundamental Concepts of Data and Knowledge > Explainable AI.

In this paper, we break with the traditional approach to classification, which is regarded as a form of supervised learning. We offer a method and algorithm, which make possible fully autonomous (unsupervised) detection of new classes, and learning following a very parsimonious training priming (few labeled data samples only). Moreover, new unknown classes may appear at a later stage and the proposed xClass method and algorithm are able to successfully discover this and learn from the data autonomously. Furthermore, the features (inputs to the classifier) are automatically sub-selected by the algorithm based on the accumulated data density per feature per class. In addition, the automatically generated model is easy to interpret and is locally generative and based on prototypes which define the modes of the data distribution. As a result, a highly efficient, lean, human-understandable, autonomously self-learning model (which only needs an extremely parsimonious priming) emerges from the data. To validate our proposal, we approbated it on four challenging problems, including imbalanced Faces-1999 data base, Caltech-101 dataset, vehicles dataset, and iRoads dataset, which is a dataset of images of autonomous driving scenarios. Not only we achieved higher precision (in one of the problems outperforming by 25% all other methods), but, more significantly, we only used a single class beforehand, while other methods used all the available classes and we generated interpretable models with smaller number of features used, through extremely weak and weak supervision. We demonstrated the ability to detect and learn new classes for both images and numerical examples.

This paper proposes a new extended zero-order Autonomous Learning Multiple-Model (ALMMo-0*) neuro-fuzzy approach in order to classify different heart disorders through sounds. ALMMo-0* is build upon the recently introduced ALMMo-0. In this paper ALMMo-0 is extended by adding a pre-processing structure which improves the the performance of the proposed method. ALMMo-0* has as a learning engine composed of hierarchical a massively parallel set of 0-order fuzzy rules, which are able to self-adapt and provide transparent and human understandable IF ... THEN representation. The heart sound recordings considered in the analysis were sourced from several contributors around the world. Data were collected from both clinical and nonclinical environment, and from healthy and pathological patients. Differently from mainstream machine learning approaches, ALMMo-0* is able to learn from unseen data. The main goal of the proposed method is to provide highly accurate models with high transparency, interpretability, and explainability for heart disorder diagnosis. Experiments demonstrated that the proposed neuro-fuzzy-based modeling is an efficient framework for these challenging classification tasks surpassing its state-of-the-art competitors in terms of classification accuracy. Additionally, ALMMo-0* produced transparent AnYa type fuzzy rules, which are human interpretable, and may help specialists to provide more accurate diagnosis. Medical doctors can easily identify abnormal heart sounds by comparing a patient's sample with the identified prototypes from abnormal samples by ALMMo-0*.

Traditionally, in supervised machine learning, (a significant) part of the available data (usually 50%-80%) is used for training and the rest-for validation. In many problems, however, the data are highly imbalanced in regard to different classes or does not have good coverage of the feasible data space which, in turn, creates problems in validation and usage phase. In this paper, we propose a technique for synthesizing feasible and likely data to help balance the classes as well as to boost the performance in terms of confusion matrix as well as overall. The idea, in a nutshell, is to synthesize data samples in close vicinity to the actual data samples specifically for the less represented (minority) classes. This has also implications to the so-called fairness of machine learning. In this paper, we propose a specific method for synthesizing data in a way to balance the classes and boost the performance, especially of the minority classes. It is generic and can be applied to different base algorithms, for example, support vector machines, k-nearest neighbour classifiers deep neural, rule-based classifiers, decision trees, and so forth. The results demonstrated that (a) a significantly more balanced (and fair) classification results can be achieved and (b) that the overall performance as well as the performance per class measured by confusion matrix can be boosted. In addition, this approach can be very valuable for the cases when the number of actual available labelled data is small which itself is one of the problems of the contemporary machine learning.

The COVID-19 disease has widely spread all over the world since the beginning of 2020. On January 30, 2020 the World Health Organization (WHO) declared a global health emergency. At the time of writing this paper the number of infected about 2 million people worldwide and took over 125,000 lives, the advanced public health systems of European countries as well as of USA were overwhelmed. In this paper, we propose an eXplainable Deep Learning approach to detect COVID-19 from computer tomography (CT) - Scan images. The rapid detection of any COVID-19 case is of supreme importance to ensure timely treatment. From a public health perspective, rapid patient isolation is also extremely important to curtail the rapid spread of the disease. From this point of view the proposed method offers an easy to use and understand tool to the front-line medics. It is of huge importance not only the statistical accuracy and other measures, but also the ability to understand and interpret how the decision was made. The results demonstrate that the proposed approach is able to surpass the other published results which were using standard Deep Neural Network in terms of performance. Moreover, it produce highly interpretable results which may be helpful for the early detection of the disease by specialists.

This paper presents an actively semi-supervised multi-layer neuro-fuzzy modeling method, ASSDRB, to classify different lighting conditions for driving scenes. ASSDRB is composed of a massively parallel ensemble of AnYa type 0-order fuzzy rules. It uses a recursive learning algorithm to update its structure when new data items are provided and, therefore, is able to cope with nonstationarities. Different lighting conditions for driving situations are considered in the analysis, which is used by self-driving cars as a safety mechanism. Differently from mainstream Deep Neural Networks approaches, the ASSDRB is able to learn from unseen data. Experiments on different lighting conditions for driving scenes, demonstrated that the deep neuro-fuzzy modeling is an efficient framework for these challenging classification tasks. Classification accuracy is higher than those produced by alternative machine learning methods. The number of algebraic calculations for the present method are significantly smaller and, therefore, the method is significantly faster than common Deep Neural Networks approaches. Moreover, DRB produced transparent AnYa fuzzy rules, which are human interpretable.

Recidivism prediction provides decision makers with an assessment of the likelihood that a criminal defendant will reoffend that can be used in pre-trial decision-making. It can also be used for prediction of locations where crimes most occur, profiles that are more likely to commit violent crimes. While such instruments are gaining increasing popularity, their use is controversial as they may present potential discriminatory bias in the risk assessment. In this paper we propose a new fair-by-design approach to predict recidivism. It is prototype-based, learns locally and extracts empirically the data distribution. The results show that the proposed method is able to reduce the bias and provide human interpretable rules to assist specialists in the explanation of the given results.

In this paper we offer a method and algorithm, which make possible fully autonomous (unsupervised) detection of new classes, and learning following a very parsimonious training priming (few labeled data samples only). Moreover, new unknown classes may appear at a later stage and the proposed xClass method and algorithm are able to successfully discover this and learn from the data autonomously. Furthermore, the features (inputs to the classifier) are automatically sub-selected by the algorithm based on the accumulated data density per feature per class. As a result, a highly efficient, lean, human-understandable, autonomously self-learning model (which only needs an extremely parsimonious priming) emerges from the data. To validate our proposal we tested it on two challenging problems, including imbalanced Caltech-101 data set and iRoads dataset. Not only we achieved higher precision, but, more significantly, we only used a single class beforehand, while other methods used all the available classes) and we generated interpretable models with smaller number of feat ures used, through extremely weak and weak supervision.

Machine learning has arisen with advanced data analytics. Many factors influence crop yield, such as soil, amount of water, climate, and genotype. Determining factors that significantly influence yield prediction and identify the most appropriate predictive methods are important in yield management. It is critical to consider and study the combination of different crop factors and their impact on the yield. The objectives of this paper are: (1) to use advanced data analytic techniques to precisely predict the soybean crop yields, (2) to identify the most influential features that impact soybean predictions, (3) to illustrate the ability of Fuzzy Rule-Based (FRB) sub-systems, which are self-organizing, self-learning, and data-driven, by using the recently developed Autonomous Learning Multiple-Model First-order (ALMMo-1) system, and (4) to compare the performance with other well-known methods. The ALMMo-1 system is a transparent model, which stakeholders can easily read and interpret. The model is a datadriven and composed of prototypes selected from the actual data. Many factors affect the yield, and data clouds can be formed in the feature/data space based on the data density. The data cloud is the key to the IF part of FRB sub-systems, while the THEN part (the consequences of the IF condition) illustrates the yield prediction in the form of a linear regression model, which consists of the yield features or factors. In addition, the model can determine the most influential features of the yield prediction online. The model shows an excellent prediction accuracy with a Root Mean Square Error (RMSE) of 0.0883, and Non-Dimensional Error Index (NDEI) of 0.0611, which is competitive with state-of-the-art methods.

This paper presents a Gaussian fuzzy set-based evolving modeling method, FBeM-G, to predict tropical cyclone tracks 6 hours in advance. FBeM-G summarizes similar data into Gaussian granules evolved from a sequence of data. It uses a recursive learning algorithm to update its parameters and structure over time and therefore is able to cope with nonstationarities. Past values of latitude, longitude, maximum sustained wind, pressure and wind radii in different quadrants of the Katrina, Sandy and Wilma tropical cyclones were obtained from the `best track' analysis provided by the National Hurricane Center (NOAA). An ensemble of cloud-based and fuzzy models was considered to compare the estimated tracks. FBeM-G provided more accurate 6-hourly track estimates using a smaller number of local models and parameters. Although less accurate, longer-term estimates given by the ensemble approach became closer to those provided by FBeM-G. An outer approximation of the pointwise track prediction is a particular characteristic of the method that is useful to determine risk areas and actions to be taken.

A computer vision approach to classify garbage into recycling categories could be an efficient way to process waste. This project aims to take garbage waste images and classify them into four classes: glass, paper, metal and, plastic. We use a garbage image database that contains around 400 images for each class. The models used in the experiments are Pre-trained VGG-16 (VGG16), AlexNet, Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and, Random Forest (RF). Experiments showed that our models reached accuracy of around 93%.

This paper concerns the application of a cloud-based intelligent evolving method, namely, a typicality-and-eccentricity-based method for data analysis (TEDA), to predict monthly mean temperature in different cities of Brazil. Past values of maximum, minimum and mean monthly temperature, as well as previous values of exogenous variables such as cloudiness, rainfall and humidity were considered in the analysis. A non-parametric Spearman correlation based method is proposed to rank and select the most relevant features and time delays for a more accurate prediction. The datasets were obtained from weather stations located in main cities such as Sao Paulo, Manaus, and Porto Alegre. These cities are known to have particular weather characteristics. TEDA prediction results are compared with results provided by the evolving Takagi-Sugeno (eTS) and the extended Takagi-Sugeno (xTS) methods. In general, TEDA provided slightly more accurate predictions at the price of a higher computational cost.