All issues

We found a predominance of α-rhythm patterns in the left hemisphere in healthy people compared to people with COVID-19 history. Moreover, we observe a significant decrease in the left hemisphere contribution to the speech center area in people who have undergone COVID-19 when performing speech tasks.

Our findings show that the signal in healthy subjects is more spatially localized and synchronized between hemispheres when performing tasks compared to people who recovered from COVID-19. We also observed a decrease in low frequencies in both hemispheres after COVID-19.

EEG-patterns of COVID-19 are detectable in an unusual frequency domain. What is usually considered noise in electroencephalographic (EEG) data carries information that can be used to determine whether or not a person has had COVID-19. These patterns can be interpreted as signs of hemispheric desynchronization, premature brain ageing, and more significant brain strain when performing simple tasks compared to people who did not have COVID-19.

In our work, we have shown the applicability of neural networks in helping to detect the long-term effects of COVID-19 on EEG-data. Furthermore, our data following other studies supported the hypothesis of the severity of the long-term effects of COVID-19 detected on the EEG-data of EEG-based BCI. The presented findings of functional activity of the brain– computer interface make it possible to use machine learning methods on simple, non-invasive brain–computer interfaces to detect post-COVID syndrome and develop progress in neurorehabilitation.

The aim of study was to develop an ensemble machine learning method for constructing interpretable predictive models and to validate it using the example of predicting in-hospital mortality (IHM) in patients with ST-segment elevation myocardial infarction (STEMI).

A retrospective cohort study was conducted using data from 5446 electronic medical records of STEMI patients who underwent percutaneous coronary intervention (PCI). Patients were divided into two groups: 335 (6.2%) patients who died during hospitalization and 5111 (93.8%) patients with a favourable in-hospital outcome. A pool of potential predictors was formed using statistical methods. Through multimetric categorization (minimizing p-values, maximizing the area under the ROC curve (AUC), and SHAP value analysis), decision trees, and multivariable logistic regression (MLR), predictors were transformed into risk factors for IHM. Predictive models for IHM were developed using MLR, Random Forest Risk Factors (RandFRF), Stochastic Gradient Boosting (XGboost), Random Forest (RF), Adaptive boosting, Gradient Boosting, Light Gradient-Boosting Machine, Categorical Boosting (CatBoost), Explainable Boosting Machine and Stacking methods.

Authors developed the RandFRF method, which integrates the predictive outcomes of modified decision trees, identifies risk factors and ranks them based on their contribution to the risk of adverse outcomes. RandFRF enables the development of predictive models with high discriminative performance (AUC 0.908), comparable to models based on CatBoost and Stacking (AUC 0.904 and 0.908, respectively). In turn, risk factors provide clinicians with information on the patient’s risk group classification and the extent of their impact on the probability of IHM. The risk factors identified by RandFRF can serve not only as rationale for the prediction results but also as a basis for developing more accurate models.

Posts on social networks can both predict the movement of the financial market, and in some cases even determine its direction. The analysis of posts on Twitter contributes to the prediction of cryptocurrency prices. The specificity of the community is represented in a special vocabulary. Thus, slang expressions and abbreviations are used in posts, the presence of which makes it difficult to vectorize text data, as a result of which preprocessing methods such as Stanza lemmatization and the use of regular expressions are considered. This paper describes created simplest machine learning models, which may work despite such problems as lack of data and short prediction timeframe. A word is considered as an element of a binary vector of a data unit in the course of the problem of binary classification solving. Basic words are determined according to the frequency analysis of mentions of a word. The markup is based on Binance candlesticks with variable parameters for a more accurate description of the trend of price changes. The paper introduces metrics that reflect the distribution of words depending on their belonging to a positive or negative classes. To solve the classification problem, we used a dense model with parameters selected by Keras Tuner, logistic regression, a random forest classifier, a naive Bayesian classifier capable of working with a small sample, which is very important for our task, and the k-nearest neighbors method. The constructed models were compared based on the accuracy metric of the predicted labels. During the investigation we recognized that the best approach is to use models which predict price movements of a single coin. Our model deals with posts that mention LUNA project, which no longer exist. This approach to solving binary classification of text data is widely used to predict the price of an asset, the trend of its movement, which is often used in automated trading.

Distributed saddle point problems (SPPs) have numerous applications in optimization, matrix games and machine learning. For example, the training of generated adversarial networks is represented as a min-max optimization problem, and training regularized linear models can be reformulated as an SPP as well. This paper studies distributed nonsmooth SPPs with Lipschitz-continuous objective functions. The objective function is represented as a sum of several components that are distributed between groups of computational nodes. The nodes, or agents, exchange information through some communication network that may be centralized or decentralized. A centralized network has a universal information aggregator (a server, or master node) that directly communicates to each of the agents and therefore can coordinate the optimization process. In a decentralized network, all the nodes are equal, the server node is not present, and each agent only communicates to its immediate neighbors.

We assume that each of the nodes locally holds its objective and can compute its value at given points, i. e. has access to zero-order oracle. Zero-order information is used when the gradient of the function is costly, not possible to compute or when the function is not differentiable. For example, in reinforcement learning one needs to generate a trajectory to evaluate the current policy. This policy evaluation process can be interpreted as the computation of the function value. We propose an approach that uses a smoothing technique, i. e., applies a first-order method to the smoothed version of the initial function. It can be shown that the stochastic gradient of the smoothed function can be viewed as a random two-point gradient approximation of the initial function. Smoothing approaches have been studied for distributed zero-order minimization, and our paper generalizes the smoothing technique on SPPs.

In this paper the various methods of machine classification of scientific texts by thematic sections on the example of publications in specialized medical journals published by Springer are compared. The corpus of texts was studied in five sections: pharmacology/toxicology, cardiology, immunology, neurology and oncology. We considered both classification methods based on the analysis of annotations and keywords, and classification methods based on the processing of actual texts. Methods of Bayesian classification, reference vectors, and reference letter combinations were applied. It is shown that the method of classification with the best accuracy is based on creating a library of standards of letter trigrams that correspond to texts of a certain subject. It is turned out that for this corpus the Bayesian method gives an error of about 20%, the support vector machine has error of order 10%, and the proximity of the distribution of three-letter text to the standard theme gives an error of about 5%, which allows to rank these methods to the use of artificial intelligence in the task of text classification by industry specialties. It is important that the support vector method provides the same accuracy when analyzing annotations as when analyzing full texts, which is important for reducing the number of operations for large text corpus.

Accurate tree identification is essential for ecological monitoring, biodiversity assessment, and forest management. Traditional manual survey methods are labor-intensive and ineffective over large areas. Advances in remote sensing technologies including lidar and hyperspectral imaging improve automated, exact detection in many fields.

Nevertheless, these technologies typically require extensive labeled data and manual feature engineering, which restrict scalability. This research proposes a new method of Self-Supervised Learning (SSL) with the SimCLR framework to enhance the classification of tree species using unlabelled data. SSL model automatically discovers strong features by merging the spectral data from hyperspectral data with the structural data from LiDAR, eliminating the need for manual intervention.

We evaluate the performance of the SSL model against traditional classifiers, including Random Forest (RF), Support Vector Machines (SVM), and Supervised Learning methods, using a dataset from the ECODSE competition, which comprises both labeled and unlabeled samples of tree species in Florida’s Ordway-Swisher Biological Station. The SSL method has been demonstrated to be significantly more effective than traditional methods, with a validation accuracy of 97.5% compared to 95.56% for Semi-SSL and 95.03% for CNN in Supervised Learning.

Subsampling experiments showed that the SSL technique is still effective with less labeled data, with the model achieving good accuracy even with only 20% labeled data points. This conclusion demonstrates SSL’s practical applications in circumstances with insufficient labeled data, such as large-scale forest monitoring.

The problem is to automate the correction of LaTeX documents. Each document is represented as a parse tree. The modified Zhang-Shasha algorithm is used to construct a mapping of tree vertices of the original document to the tree vertices of the edited document, which corresponds to the minimum editing distance. Vertex to vertex maps form the training set, which is used to generate rules for automatic correction. The statistics of the applicability to the edited documents is collected for each rule. It is used for quality assessment and improvement of the rules.

Most biomechanical tasks of interest to clinicians can be solved only using personalized mathematical models. Such models allow to formalize and relate key pathophysiological processes, basing on clinically available data evaluate non-measurable parameters that are important for the diagnosis of diseases, predict the result of a therapeutic or surgical intervention. The use of models in clinical practice imposes additional restrictions: clinicians require model validation on clinical cases, the speed and automation of the entire calculated technological chain, from processing input data to obtaining a result. Limitations on the simulation time, determined by the time of making a medical decision (of the order of several minutes), imply the use of reduction methods that correctly describe the processes under study within the framework of reduced models or machine learning tools.

Personalization of models requires patient-oriented parameters, personalized geometry of a computational domain and generation of a computational mesh. Model parameters are estimated by direct measurements, or methods of solving inverse problems, or methods of machine learning. The requirement of personalization imposes severe restrictions on the number of fitted parameters that can be measured under standard clinical conditions. In addition to parameters, the model operates with boundary conditions that must take into account the patient’s characteristics. Methods for setting personalized boundary conditions significantly depend on the clinical setting of the problem and clinical data. Building a personalized computational domain through segmentation of medical images and generation of the computational grid, as a rule, takes a lot of time and effort due to manual or semi-automatic operations. Development of automated methods for setting personalized boundary conditions and segmentation of medical images with the subsequent construction of a computational grid is the key to the widespread use of mathematical modeling in clinical practice.

The aim of this work is to review our solutions for personalization of mathematical models within the framework of three tasks of clinical cardiology: virtual assessment of hemodynamic significance of coronary artery stenosis, calculation of global blood flow after hemodynamic correction of complex heart defects, calculating characteristics of coaptation of reconstructed aortic valve.

The purpose of this study is to develop a methodology for change points detection in time series, including financial data. The theoretical basis of the study is based on the pieces of research devoted to the analysis of structural changes in financial markets, description of the proposed algorithms for detecting change points and peculiarities of building classical and deep machine learning models for solving this type of problems. The development of such tools is of interest to investors and other stakeholders, providing them with additional approaches to the effective analysis of financial markets and interpretation of available data.

To address the research objective, a neural network was trained. In the course of the study several ways of training sample formation were considered, differing in the nature of statistical parameters. In order to improve the quality of training and obtain more accurate results, a methodology for feature generation was developed for the formation of features that serve as input data for the neural network. These features, in turn, were derived from an analysis of mathematical expectations and standard deviations of time series data over specific intervals. The potential for combining these features to achieve more stable results is also under investigation.

The results of model experiments were analyzed to compare the effectiveness of the proposed model with other existing changepoint detection algorithms that have gained widespread usage in practical applications. A specially generated dataset, developed using proprietary methods, was utilized as both training and testing data. Furthermore, the model, trained on various features, was tested on daily data from the S&P 500 index to assess its effectiveness in a real financial context.

As the principles of the model’s operation are described, possibilities for its further improvement are considered, including the modernization of the proposed model’s structure, optimization of training data generation, and feature formation. Additionally, the authors are tasked with advancing existing concepts for real-time changepoint detection.

Disassembly of stripped x86 binaries is an important yet non-trivial task. Disassembly is difficult to perform correctly without debug information, especially on x86 architecture, which has variablesized instructions interleaved with data. Moreover, the presence of indirect jumps in binary code adds another layer of complexity. Indirect jumps impede the ability of recursive traversal, a common disassembly technique, to successfully identify all instructions within the code. Consequently, disassembling such code becomes even more intricate and demanding, further highlighting the challenges faced in this field. Many tools, including commercial ones such as IDA Pro, struggle with accurate x86 disassembly. As such, there has been some interest in developing a better solution using machine learning (ML) techniques. ML can potentially capture underlying compiler-independent patterns inherent for the compiler-generated assembly. Researchers in this area have shown that it is possible for ML approaches to outperform the classical tools. They also can be less timeconsuming to develop compared to manual heuristics, shifting most of the burden onto collecting a big representative dataset of executables with debug information. Following this line of work, we propose an improvement of an existing RGCN-based architecture, which builds control and flow graph on superset disassembly. The enhancement comes from augmenting the graph with data flow information. In particular, in the embedding we add Jump Control Flow and Register Dependency edges, inspired by Probabilistic Disassembly. We also create an open-source x86 instruction identification dataset, based on a combination of ByteWeight dataset and a selection open-source Debian packages. Compared to IDA Pro, a state of the art commercial tool, our approach yields better accuracy, while maintaining great performance on our benchmarks. It also fares well against existing machine learning approaches such as DeepDi.