Deep Learning Hybrid Techniques for Brain Tumor Segmentation.

Medical images play an important role in medical diagnosis and treatment. Oncologists analyze images to determine the different characteristics of deadly diseases, plan the therapy, and observe the evolution of the disease. The objective of this paper is to propose a method for the detection of brain tumors. Brain tumors are identified from Magnetic Resonance (MR) images by performing suitable segmentation procedures. The latest technical literature concerning radiographic images of the brain shows that deep learning methods can be implemented to extract specific features of brain tumors, aiding clinical diagnosis. For this reason, most data scientists and AI researchers work on Machine Learning methods for designing automatic screening procedures. Indeed, an automated method would result in quicker segmentation findings, providing a robust output with respect to possible differences in data sources, mostly due to different procedures in data recording and storing, resulting in a more consistent identification of brain tumors. To improve the performance of the segmentation procedure, new architectures are proposed and tested in this paper. We propose deep neural networks for the detection of brain tumors, trained on the MRI scans of patients' brains. The proposed architectures are based on convolutional neural networks and inception modules for brain tumor segmentation. A comparison of these proposed architectures with the baseline reference ones shows very interesting results. MI-Unet showed a performance increase in comparison to baseline Unet architecture by 7.5% in dice score, 23.91% insensitivity, and 7.09% in specificity. Depth-wise separable MI-Unet showed a performance increase by 10.83% in dice score, 2.97% in sensitivity, and 12.72% in specificity as compared to the baseline Unet architecture. Hybrid Unet architecture achieved performance improvement of 9.71% in dice score, 3.56% in sensitivity, and 12.6% in specificity. Whereas the depth-wise separable hybrid Unet architecture outperformed the baseline architecture by 15.45% in dice score, 20.56% in sensitivity, and 12.22% in specificity.

(Hyper)graph Kernels over Simplicial Complexes.

Graph kernels are one of the mainstream approaches when dealing with measuring similarity between graphs, especially for pattern recognition and machine learning tasks. In turn, graphs gained a lot of attention due to their modeling capabilities for several real-world phenomena ranging from bioinformatics to social network analysis. However, the attention has been recently moved towards hypergraphs, generalization of plain graphs where multi-way relations (other than pairwise relations) can be considered. In this paper, four (hyper)graph kernels are proposed and their efficiency and effectiveness are compared in a twofold fashion. First, by inferring the simplicial complexes on the top of underlying graphs and by performing a comparison among 18 benchmark datasets against state-of-the-art approaches; second, by facing a real-world case study (i.e., metabolic pathways classification) where input data are natively represented by hypergraphs. With this work, we aim at fostering the extension of graph kernels towards hypergraphs and, more in general, bridging the gap between structural pattern recognition and the domain of hypergraphs.

Modelling and Recognition of Protein Contact Networks by Multiple Kernel Learning and Dissimilarity Representations.

Multiple kernel learning is a paradigm which employs a properly constructed chain of kernel functions able to simultaneously analyse different data or different representations of the same data. In this paper, we propose an hybrid classification system based on a linear combination of multiple kernels defined over multiple dissimilarity spaces. The core of the training procedure is the joint optimisation of kernel weights and representatives selection in the dissimilarity spaces. This equips the system with a two-fold knowledge discovery phase: by analysing the weights, it is possible to check which representations are more suitable for solving the classification problem, whereas the pivotal patterns selected as representatives can give further insights on the modelled system, possibly with the help of field-experts. The proposed classification system is tested on real proteomic data in order to predict proteins' functional role starting from their folded structure: specifically, a set of eight representations are drawn from the graph-based protein folded description. The proposed multiple kernel-based system has also been benchmarked against a clustering-based classification system also able to exploit multiple dissimilarities simultaneously. Computational results show remarkable classification capabilities and the knowledge discovery analysis is in line with current biological knowledge, suggesting the reliability of the proposed system.

Modeling failures in smart grids by a bilinear logistic regression approach.

Modeling and recognizing events in complex systems through machine learning techniques is a challenging task. Especially if the model is constrained to be explainable and interpretable, while ensuring high levels of accuracy. In this paper, we adopt a bilinear logistic regression model in which the parameters are trained in a data-driven fashion on a real-world dataset of power grid failure data. The bilinear white-box model - grounded on a specific neural architecture - has been proven effective in classifying faulty states with a performance comparable to several classifiers in technical literature. Additionally, the low computational complexity of the bilinear model, in terms of the number of free parameters, allows gaining insights into the fault phenomenon correlating the events that impact the power grid (exogenous causes) with its constitutive characteristics, thence eliciting the relational information hidden in the data. The proposed model is also able to estimate a vulnerability vector that can be associated, as a suitable characteristic "label", to power grid components, opening the way, as will be deeply demonstrated in the following, not only to predictive maintenance programs or condition monitoring tasks but also to risk assessment and scenario analyses in line with the explainable AI paradigm.

Human Versus Machine Intelligence: Assessing Natural Language Generation Models Through Complex Systems Theory.

The introduction of Transformer architectures - with the self-attention mechanism - in automatic Natural Language Generation (NLG) is a breakthrough in solving general task-oriented problems, such as the simple production of long text excerpts that resemble ones written by humans. While the performance of GPT-X architectures is there for all to see, many efforts are underway to penetrate the secrets of these black-boxes in terms of intelligent information processing whose output statistical distributions resemble that of natural language. In this work, through the complexity science framework, a comparative study of the stochastic processes underlying the texts produced by the English version of GPT-2 with respect to texts produced by human beings, notably novels in English and programming codes, is offered. The investigation, of a methodological nature, consists first of all of an analysis phase in which the Multifractal Detrended Fluctuation Analysis and the Recurrence Quantification Analysis - together with Zipf's law and approximate entropy - are adopted to characterize long-term correlations, regularities and recurrences in human and machine-produced texts. Results show several peculiarities and trends in terms of long-range correlations and recurrences in the last case. The synthesis phase, on the other hand, uses the complexity measures to build synthetic text descriptors - hence a suitable text embedding - which serve to constitute the features for feeding a machine learning system designed to operate feature selection through an evolutionary technique. Using multivariate analysis, it is then shown the grouping tendency of the three analyzed text types, allowing to place GTP-2 texts in between natural language texts and computer codes. Similarly, the classification task demonstrates that, given the high accuracy obtained in the automatic discrimination of text classes, the proposed set of complexity measures is highly informative. These interesting results allow us to add another piece to the theoretical understanding of the surprising results obtained by NLG systems based on deep learning and let us to improve the design of new informetrics or text mining systems for text classification, fake news detection, or even plagiarism detection.

Multifractal Characterization of Texts for Pattern Recognition: On the Complexity of Morphological Structures in Modern and Ancient Languages.

The study of languages' structure and their organization in a set of well-defined relation schemes is a delicate matter. In the last decades, the convergence of traditional conflicting views by linguists is supported by an interdisciplinary approach that involves not only genetics or bio-archelogy but nowadays even the science of complexity. In light of this new and useful approach, this study proposes an in-depth analysis of the complexity underlying the morphological organization, in terms of multifractality and long-range correlations, of several modern and ancient texts pertaining to various linguistic strains (including ancient Greek, Arabic, Coptic, Neo-Latin and Germanic languages). The methodology is grounded on the mapping procedure between lexical categories belonging to text excerpts and time series, which is based on the rank of the frequency occurrence. Through the well-known MFDFA technique and a specific multifractal formalism, several multifractal indexes are then extracted for characterizing texts and the multifractal signature has been adopted for characterizing several language families, such as Indo-European, Semitic and Hamito-Semitic. The regularities and differences in the linguistic strains are assessed within a multivariate statistical framework and corroborated with a Machine Learning approach that is dedicated, in turn, to investigate the predictive power of the multifractal signature pertinent to text excerpts. The obtained results show a strong presence of persistence, i.e., memory, in the morphological structure of analyzed texts and we claim that this property has a role in characterizing the studied linguistic families. In fact, for example, the proposed analysis framework - grounded on complexity indexes - is able to easily distinguish ancient Greek texts from Arabic ones, as they belong to different language strains, i.e., indo-European and Semitic, respectively. The proposed approach has been proven effective and can be adopted for further comparative studies and for designing new informetrics for further advances in the fields of information retrieval and Artificial Intelligence.

On component-wise dissimilarity measures and metric properties in pattern recognition.

In many real-world applications concerning pattern recognition techniques, it is of utmost importance the automatic learning of the most appropriate dissimilarity measure to be used in object comparison. Real-world objects are often complex entities and need a specific representation grounded on a composition of different heterogeneous features, leading to a non-metric starting space where Machine Learning algorithms operate. However, in the so-called unconventional spaces a family of dissimilarity measures can be still exploited, that is, the set of component-wise dissimilarity measures, in which each component is treated with a specific sub-dissimilarity that depends on the nature of the data at hand. These dissimilarities are likely to be non-Euclidean, hence the underlying dissimilarity matrix is not isometrically embeddable in a standard Euclidean space because it may not be structurally rich enough. On the other hand, in many metric learning problems, a component-wise dissimilarity measure can be defined as a weighted linear convex combination and weights can be suitably learned. This article, after introducing some hints on the relation between distances and the metric learning paradigm, provides a discussion along with some experiments on how weights, intended as mathematical operators, interact with the Euclidean behavior of dissimilarity matrices.

An Infoveillance System for Detecting and Tracking Relevant Topics From Italian Tweets During the COVID-19 Event.

The year 2020 opened with a dramatic epidemic caused by a new species of coronavirus that soon has been declared a pandemic by the WHO due to the high number of deaths and the critical mass of worldwide hospitalized patients, of order of millions. The COVID-19 pandemic has forced the governments of hundreds of countries to apply several heavy restrictions in the citizens' socio-economic life. Italy was one of the most affected countries with long-term restrictions, impacting the socio-economic tissue. During this lockdown period, people got informed mostly on Online Social Media, where a heated debate followed all main ongoing events. In this scenario, the following study presents an in-depth analysis of the main emergent topics discussed during the lockdown phase within the Italian Twitter community. The analysis has been conducted through a general purpose methodological framework, grounded on a biological metaphor and on a chain of NLP and graph analysis techniques, in charge of detecting and tracking emerging topics in Online Social Media, e.g. streams of Twitter data. A term-frequency analysis in subsequent time slots is pipelined with nutrition and energy metrics for computing hot terms by also exploiting the tweets quality information, such as the social influence of the users. Finally, a co-occurrence analysis is adopted for building a topic graph where emerging topics are suitably selected. We demonstrate via a careful parameter setting the effectiveness of the topic tracking system, tailored to the current Twitter standard API restrictions, in capturing the main sociopolitical events that occurred during this dramatic phase.

Metabolic networks classification and knowledge discovery by information granulation.

Graphs are powerful structures able to capture topological and semantic information from data, hence suitable for modelling a plethora of real-world (complex) systems. For this reason, graph-based pattern recognition gained a lot of attention in recent years. In this paper, a general-purpose classification system in the graphs domain is presented. When most of the information of the available patterns can be encoded in edge labels, an information granulation-based approach is highly discriminant and allows for the identification of semantically meaningful edges. The proposed classification system has been tested on the entire set of organisms (5299) for which metabolic networks are known, allowing for both a perfect mirroring of the underlying taxonomy and the identification of most discriminant metabolic reactions and pathways. The widespread diffusion of graph (network) structures in biology makes the proposed pattern recognition approach potentially very useful in many different fields of application. More specifically, the possibility to have a reliable metric to compare different metabolic systems is instrumental in emerging fields like microbiome analysis and, more in general, for proposing metabolic networks as a universal phenotype spanning the entire tree of life and in direct contact with environmental cues.

A White-Box Equivalent Neural Network Circuit Model for SoC Estimation of Electrochemical Cells.

Smart grids, microgrids, and pure electric powertrains are the key technologies for achieving the expected goals concerning the restraint of CO2 emissions and global warming. In this context, an effective use of electrochemical energy storage systems (ESSs) is mandatory. In particular, accurate state of charge (SoC) estimations are helpful for improving the ESS performances. To this aim, developing accurate models of electrochemical cells is necessary for implementing effective SoC estimators. Therefore, a novel neural network modeling technique is proposed in this paper. The main contribution consists in the development of a white-box neural design that provides helpful insights into the cell physics, together with a powerful nonlinear approximation capability, and a flexible system identification procedure. In order to do that, the system equations of a white-box equivalent circuit model (ECM) have been combined with computational intelligence techniques by approximating each circuit element with a dedicated neural network. The model performances have been analyzed in terms of model accuracy, SoC estimation effectiveness, and computational cost over two realistic data sets. Moreover, the proposed model has been compared with a white-box ECM and a gray-box neural network model. The results prove that the proposed modeling technique is able to provide useful improvements in the SoC estimation task with a competing computational cost.

Cancer Diagnosis Using Deep Learning: A Bibliographic Review.

In this paper, we first describe the basics of the field of cancer diagnosis, which includes steps of cancer diagnosis followed by the typical classification methods used by doctors, providing a historical idea of cancer classification techniques to the readers. These methods include Asymmetry, Border, Color and Diameter (ABCD) method, seven-point detection method, Menzies method, and pattern analysis. They are used regularly by doctors for cancer diagnosis, although they are not considered very efficient for obtaining better performance. Moreover, considering all types of audience, the basic evaluation criteria are also discussed. The criteria include the receiver operating characteristic curve (ROC curve), Area under the ROC curve (AUC), F1 score, accuracy, specificity, sensitivity, precision, dice-coefficient, average accuracy, and Jaccard index. Previously used methods are considered inefficient, asking for better and smarter methods for cancer diagnosis. Artificial intelligence and cancer diagnosis are gaining attention as a way to define better diagnostic tools. In particular, deep neural networks can be successfully used for intelligent image analysis. The basic framework of how this machine learning works on medical imaging is provided in this study, i.e., pre-processing, image segmentation and post-processing. The second part of this manuscript describes the different deep learning techniques, such as convolutional neural networks (CNNs), generative adversarial models (GANs), deep autoencoders (DANs), restricted Boltzmann's machine (RBM), stacked autoencoders (SAE), convolutional autoencoders (CAE), recurrent neural networks (RNNs), long short-term memory (LTSM), multi-scale convolutional neural network (M-CNN), multi-instance learning convolutional neural network (MIL-CNN). For each technique, we provide Python codes, to allow interested readers to experiment with the cited algorithms on their own diagnostic problems. The third part of this manuscript compiles the successfully applied deep learning models for different types of cancers. Considering the length of the manuscript, we restrict ourselves to the discussion of breast cancer, lung cancer, brain cancer, and skin cancer. The purpose of this bibliographic review is to provide researchers opting to work in implementing deep learning and artificial neural networks for cancer diagnosis a knowledge from scratch of the state-of-the-art achievements.

A Novel Neural Networks Ensemble Approach for Modeling Electrochemical Cells.

Accurate modeling of electrochemical cells is nowadays mandatory for achieving effective upgrades in the fields of energetic efficiency and sustainable mobility. Indeed, these models are often used for performing accurate State-of-Charge (SoC) estimations in energy storage systems used in microgrids or powering pure electric and hybrid cars. To this aim, a novel neural networks ensemble approach for modeling electrochemical cells is proposed in this paper. Herein, the system identification has been faced by means of a gray box technique, in which different and specialized neural networks are used for identifying the unknown internal behaviors of the cell. In particular, the a priori knowledge on the system dynamic is used for defining the network architecture. Specifically, each nonlinear function appearing in the system equations is approximated by a distinct neural network. The proposed model has been validated upon three different data sets both in terms of model accuracy and effectiveness in the SoC estimation task. The achieved performances have been compared with those of other computational intelligence approaches proposed in the literature. The results prove the effectiveness of the gray box scheme, achieving very promising performances in both the system identification accuracy and the SoC estimation task.

A generative model for protein contact networks.

In this paper, we present a generative model for protein contact networks (PCNs). The soundness of the proposed model is investigated by focusing primarily on mesoscopic properties elaborated from the spectra of the graph Laplacian. To complement the analysis, we also study the classical topological descriptors, such as statistics of the shortest paths and the important feature of modularity. Our experiments show that the proposed model results in a considerable improvement with respect to two suitably chosen generative mechanisms, mimicking with better approximation real PCNs in terms of diffusion properties elaborated from the normalized Laplacian spectra. However, as well as the other network models, it does not reproduce with sufficient accuracy the shortest paths structure. To compensate this drawback, we designed a second step involving a targeted edge reconfiguration process. The ensemble of reconfigured networks denotes further improvements that are statistically significant. As an important byproduct of our study, we demonstrate that modularity, a well-known property of proteins, does not entirely explain the actual network architecture characterizing PCNs. In fact, we conclude that modularity, intended as a quantification of an underlying community structure, should be considered as an emergent property of the structural organization of proteins. Interestingly, such a property is suitably optimized in PCNs together with the feature of path efficiency.

Prediction of telephone calls load using Echo State Network with exogenous variables.

We approach the problem of forecasting the load of incoming calls in a cell of a mobile network using Echo State Networks. With respect to previous approaches to the problem, we consider the inclusion of additional telephone records regarding the activity registered in the cell as exogenous variables, by investigating their usefulness in the forecasting task. Additionally, we analyze different methodologies for training the readout of the network, including two novel variants, namely ν-SVR and an elastic net penalty. Finally, we employ a genetic algorithm for both the tasks of tuning the parameters of the system and for selecting the optimal subset of most informative additional time-series to be considered as external inputs in the forecasting problem. We compare the performances with standard prediction models and we evaluate the results according to the specific properties of the considered time-series.