AUTOMATIC DOCUMENTATION AND MATHEMATICAL LINGUISTICS最新文献

The paper analyzes the limitations of conventional methods for assessing the maturity of technology, such as the S-curve, technology readiness level (TRL), Gartner’s hype cycle and their dependence on experts’ opinions. Current approaches to this task based on big text data analysis and machine learning algorithms are reviewed, and their advantages are demonstrated. As a result of existing research systematization, the prospects of transition to automated technology maturity assessment using machine learning methods are revealed.

This article explores the unification of two relevant areas that are rarely studied together: the computational linguistics field of sentiment analysis and the physico-mathematical field of entropy-complexity. Based on keywords from articles on both topics, links are established that form a single semantic space at the junction of these subject areas. The information system of the data visualization tool Semograph was used to visualize the data and identify semantic links between semantic components and between sentiment analysis and entropy complexity.

This paper describes the development of an intelligent system for detecting environmental problems based on multispectral satellite images. The functionality of the system is based on machine learning algorithms and provides the ability to highlight areas of environmental problems based on the analysis of each individual pixel of the image. The article describes the system architecture, data collection and preprocessing process, model development and training. The evaluation of the results of the system is based on real data. The effectiveness of application of the developments obtained as a result of the study for the tasks of monitoring the environmental state of the Earth’s surface for large territories is confirmed. The use of software modules provides the ability to respond quickly to emerging environmental abnormal situations.

The use of neural network language models to increase the efficiency of processing scientific and technical information (STI) on the Internet from the point of view of cognitive information retrieval is considered. A conceptual model is proposed to identify subject areas that corresponding to the object of information need based on existing STI classification systems. The model is based on the principles of coordinate indexing incorporating methods of detecting and tracking topics and the use of large language models to analyze the information space of the user’s information needs, metasearch of Internet resources, automatic abstracting and ranking of relevant documents.

The paper considers the features of an algorithm for finding a solution in the formalism of functional neural networks (FN-networks), where the knowledge base and description of the initial problem contain concepts that are defined by partially defined FN-networks with a regular structure, in the form of enumerations of sets of the same type of objects and fragments of the same structure, when their number is unknown and possibly infinite.

Structural, target, and resource conflict situations arising in management of organizational systems and algorithms that help to identify and resolve these situations are considered. This study is carried out in the interest of creating computer technologies for the intellectual support of decision-making in management in organizational systems on the basis of the provisions of system analysis, the theory of management and decision-making, and the theory of active systems and the concept of artificial intelligence.

The problem of generating effective configurations of structural-functional models of process activity using event-driven process methodology notation is discussed. A symbolic encoding system for topologies and a random search method, which allows generating many solutions based on previous options, are proposed. The presented experiments show that the proposed method helps to find effective configurations of the organization of processes that take into account the delimitations; if these delimitations cannot be implemented, then they suggest options in breach of them and introduce new options not previously proposed by experts, `comparing them according to performance evaluation data.

The conceptual framework and building of the ontology of a selected subject area is considered. The subject area is the development of engineering and computational architectures for the specific object of warehousing logistics. The research methodology is presented iteratively and includes a definition of the ontology structure and its requirements, a specification of terms, and a class hierarchy. The conceptual representation framework takes into account the distinction among objects of links identified from the data concerning the generative operation of a multimodal transport and logistics center. The chosen approach to ontology development allows the formulation of rules and operations for transforming the class hierarchy. This approach involves creating a hierarchy, where each concept can inherit attributes and relationships from a parent class, which simplifies data management within the subject area. The ontology is developed in the Protégé editor, the links expressing object relationships are defined, and a basis is created for searching axioms of dependencies among component objects of the IT architecture of the multimodal transport and logistics center. The ontology is intended to provide methodological support for decisions concerning configuring and parameterizing software and network components of the engineering and computing architectures of the multimodal transport and logistics center.

The role of information support in scientific and technical activity is indicated. A model of information resource replenishment of an automated information provision system is developed and implemented in the form of a system of models with its process and simulation. An approach of bringing a process model to the simulation is proposed, and the distribution dependencies of the demand dwell time in the mass service system are obtained. The information resource replenishment model for a typical automated information provisioning system is investigated.

A mathematical model of the subsystem of security in an automated information system with a special purpose, hierarchically decomposed by security levels, is considered, on the basis of which it is possible to develop and analyze relatively flexible rules of differentiation of access to information with high guarantee of their fulfillment with the help of the apparatus of finite nonlinear lattices. Two finite nonlinear lattices suitable for this purpose are proposed. For each, carrier, partial ordering, and algebraic operations are defined. Interpretations of these lattices are given as vector security levels, in contrast to the well-known interpretation of finite linear lattices as numerical security levels. For vector security levels, the “no read up” and “no write down” rules of the mandated security policy are defined. The equivalence of the modeling capabilities of these lattices is justified. The developed model integrates the principles of mandating and discretionary access control methods, taking subjects in the role of intermediaries between users and objects. The Bell-LaPadula model, as adapted to the formalization of security levels by finite nonlinear lattices is presented.