Системная Информатика, № 16

Download

Method of paradigmatic analysis of programming languages

The purpose of the article is to describe the method of comparison of programming languages, convenient for assessing the expressive power of languages and the complexity of the programming systems. The method is adapted to substantiate practical, objective criteria of program decomposition, which can be considered as an approach to solving the problem of factorization of very complicated definitions of programming languages and their support systems. In addition, the article presents the results of the analysis of the most well-known programming paradigms and outlines an approach to navigation in the modern expanding space of programming languages, based on the classification of paradigms on the peculiarities of problem statements and semantic characteristics of programming languages and systems with an emphasis on the criteria for the quality of programs and priorities in decision-making in their implementation. The concept of "programming paradigm" is manifested as the way of thinking in the programming process. The author thanks the organizers and participants of the conferences "Scientific Service in the Internet Environment" (http://agora.guru.ru/display.php?conf=abrau2020&page=subjects&PHPSESSID=qbn3kbhgnk8b6a9g21qi1nkkq2 ), discussions with which made it possible to understand the main provisions of this article.
Download

Verification of a predicate heapsort program using inverse transformations

Deductive verification of the classical J.Williams heapsort algorithm for objects of an arbitrary type was conducted. In order to simplify verification, non-trivial transformations, replacing pointer arithmetic operators by an array element constructs, were applied. The program was translated to the predicate programming language. Deductive verification of the program in the tools Why3 and Coq appears to be complicated and time consuming.
Download

Transformation, specification, and verification of the program calculating the elements number of a set presented by a bit vector

Transformations eliminating pointers in the memweight function in OS Linux kernel library is described. Next, the function is translated to the predicate programming language P. For the obtained predicate program, deductive verification in the Why3 tool was performed. In order to simplify verification, the program model of calculating program inner state was constructed.
Download

Named entity recognition in texts of administrative documents with deep neural networks

Named Entity Extraction (NER) is the task of extracting information from text data that belongs to predefined categories, such as organizations names, place names, people's names, etc. Within the framework of the presented work, was developed an approach for the additional training of deep neural networks with the attention mechanism (BERT architecture). It is shown that the preliminary training of the language model in the tasks of recovering the masked word and determining the semantic relatedness of two sentences can significantly improve the quality of solving the problem of NER. One of the best results has been achieved in the task of extracting named entities on the RuREBus dataset. One of the key features of the described solution is the closeness of the formulation to real business problems and the selection of entities not of a general nature, but specific to the economic industry.
Download

Identification of argumentative relations in popular science texts

The presented work describes the analysis of argumentative statements included into the same text topic fragment as a recognition feature in terms of its efficiency. This study is performed with the purpose of using this feature in automatic recognition of argumentative structures presented in the popular science texts written in Russian. The topic model of a text is constructed based on superphrasal units (text fragments united by one topic) that are identified by detecting clusters of words and word-combinations with the use of scan statistics. Potential relations, extracted from topic models, are verified through the use of texts with manually annotated argumentation structures. The comparison between potential (based on topic models) and manually constructed relations is performed automatically. Macro-average scores of precision and recall are equal to 48.6% and 76.2% correspondingly.