AUTOMATIC DOCUMENTATION AND MATHEMATICAL LINGUISTICS最新文献

This paper examines issues related to the presentation of information concerning publications by Kazan Federal University researchers, teachers, graduate students, and students, as well as concerning the university’s scientific sources in the information and analytical materials of the Scilit system. Specific examples present the advantages of the complete and correct setting of metadata for scientific publications, as well as problems that arise when handling bibliographic information carelessly.

The development of an application that is based on machine learning methods for automatic annotation of web pages according to the Microdata standard is described, with the possibility of an extension to other standards and injecting data to JSX files. Datasets were collected and prepared for training machine learning (ML) models. The ML model metrics were collected and analyzed.

Alive publications are exemplars of a new genre for presenting the results of scientific research; in this type, scientific work is published online and then is constantly developed and improved by its author. Serious errors and typos are no longer fatal, nor do they haunt the author for the rest of his or her life. The reader of an alive publication knows that the author is constantly monitoring changes in this branch of science. A Russian author who supports an alive publication is currently hopelessly losing in many bibliometric indicators favored by conservative science officials. An alive publication encourages the development of the bibliographical apparatus. Each citation will soon have to be continually updated to the date of the last revision of the alive publication. It is to be expected that alive publications will spread over to the scientific world, and the author’s concern for the publication’s evolution will become like a parent’s care for the development of a child, and, to the delight of the reader, the internet will be filled with scientific publications that do not lose their relevance over time.

This paper provides an overview of the modern technologies used in processing e-mail messages to address the problem of receiving trusted e-mail and describes them. Recommended settings for successful operation are provided.

The ontology-based approach to multiagent modeling involves the implementation of a modeling system through the creation of ontologies. The IEEE 1516 Standard for Modeling and Simulation High Level Architecture is an example of a holistic implementation of an ontology-based approach to agent-based modeling. This work is devoted to a multiagent modeling system designed for modeling complex radio engineering systems (especially radar systems). This is relevant due to the need to replace part of the field tests of radio engineering systems with simulation experiments. One motivation for switching to the IEEE 1516 standard for a heavy multiagent modeling system, among others, is to ensure scalability, openness, and the multiple reuse of the developed agent models, which is completely logical, based on the existing well-developed and proven standard establishing rules for the interaction of models and the development of software interfaces. The general principles of the construction and architecture of the modeling system are given. The basic requirements for the main modeling agents and their role and place in the complex modeling system are shown, a special place among which is occupied by the simulator of the background-target environment. The possibility of combining two simulation schemes is also discussed: discrete-event and step-by-step. The fact is that the step-by-step scheme has the advantages of simplicity and clarity, and it is convenient to model processing algorithms and components of radio engineering systems. However, it is impossible to implement true autonomy and asynchrony of agents in it. Combining two modeling schemes allows you to combine their advantages.

Problems that arise in the field of traffic are of great importance. To solve existing problems, various intelligent systems are being developed, one of which is the Smart City system. This work is devoted to the development of an information and analytical system (IAS) for controlling an intelligent traffic light. The presented system consists of two levels, each of which contains a set of specific operations. The first level is responsible for detecting objects, in particular pedestrians and cars at an intersection, and the second level calculates the operating time of traffic light signals for the control signal that is transmitted to the device. For comparative analysis, the combined method histogram of oriented gradients + support vector machines (HOG+SVM). HOG was chosen, based upon counting the number of gradient directions on individual image areas, and SVM were used to construct hyperplanes in n-dimensional space to separate objects belonging to different classes. The results of an experimental study, om which the recognition of objects in images was carried out, showed the superiority of the developed information and analytical system over existing methods. The average accuracy of detecting pedestrians and cars through the IAS was 69.4%. In addition, the experiment showed that the accuracy of detecting objects in images is directly proportional to the distance from the video camera to the object.

This article considers an approach to creating a family of lightweight grammars for the Go language, with a special symbol Any denoting the skipped part of the program [1]. A formal definition of a more granular grammar is given, along with examples illustrating the increase in grammar rule granularity. The efficiency of the constructed lightweight parsers is analyzed in terms of memory usage and runtime on seven industrial repositories. It is shown that increasing the granularity of the grammar does not lead to a significant increase in parser resource consumption and varies only slightly depending on repository type and coding style in Go. Furthermore, the advantages of using lightweight grammars with Any over full grammars are summarized. An example of using a lightweight grammar to determine code complexity is presented. The results can also be applied to estimate the parser’s contribution to overall resource consumption, e.g., in tasks such as code binding and project markup.

The most common approach to creating an HTML version of a journal article used by scientific publishers is to first create an XML version of the article in accordance with the NISO Journal Article Tag Suite (JATS) standard, followed by automatic conversion to HTML and PDF. However, obtaining an XML version from a manuscript in the MS Word .docx format, often used by authors, when it contains a large number of complex formulas and tables, is a difficult task. Existing software either cannot cope with it in full or is expensive and inaccessible to small publishers on a limited budget. This paper proposes an approach to creating an HTML version of a journal article from a manuscript in .docx format containing formulas in MathType format, which does not require significant financial or time costs from the publisher. It also describes a prototype converter from .docx format to HTML and JATS XML formats that implements this approach and applicable to KIAM preprints.

The article is devoted to the development of solutions in the project of a simulator for teaching programming, intended for initial familiarization with the basic concepts of process interaction and calculation management. In the transition to multiprocessor architectures, the relevance of a special language and information support for the introduction to programming increases is growing. No matter how complex the world of parallelism is, a programmer training system will have to master it and create a methodology for fully familiarizing itself with its nonobvious phenomena. This is a sufficient reason for developing an educational programming language aimed at the initial training of primary and secondary school students, as well as junior students and nonprofessionals, for operating interacting processes and programming parallel computations. The given language has been developed through many years of experience in managing the interaction of toy robots moving on a checkered board. The material of this article material is of interest to programmers, students, and graduate students specializing in the field of systems and theoretical programming and to all those interested in the problems of modern computer science, programming, and information technology, especially the problems of parallel computing, supercomputers and the use of multiprocessor complexes and computer networks in general.

This paper addresses the automatic annotation of training datasets in the field of computer vision using machine learning methods. Data annotation is a key stage in the development and training of deep learning models, but creating labeled data often requires significant time and labor. This paper proposes a mechanism for automatic annotation based on the use of convolutional neural networks and active learning methods. The proposed methodology includes the analysis and evaluation of existing approaches to automatic annotation. The effectiveness of the proposed solutions is assessed using publicly available datasets. The results demonstrate that the proposed method significantly reduces the time required for data annotation, although operator intervention is still necessary. The literature review presents an analysis of modern annotation methods and existing automatic systems, providing a better understanding of the context and advantages of the proposed approach. The conclusion discusses the study achievements, its limitations, and possible directions for future research in this field.