This article describes the possibility of creating reflexive intellectual systems for processing natural language information based on the mathematical apparatus of the theory of non-force interaction and the prospects of building information and communication systems and technologies for managing the quality of education based on them. Tasks that can be solved by such systems are classified. It was determined that the proposed model differs from the existing ones by identifying differences in the conditional probabilities of the appearance of message addressees as the amount of informational influence of fragments of natural language text, which made it possible to automatically determine the most likely addressees of this text. The developed technology meets the requirements for simplicity, is informative and easy to understand, and is minimally expensive. The authors share the experience gained during the creation of this kind of technology and show its advantages both for computational linguistics and for information and communication systems and technologies for quality management of education. Information Technology, natural language text, reflexive approach, methods of information management, Information and Communication Systems and Technologies COLINS-2023: 7th International Conference on Computational Linguistics and Intelligent Systems, April 20-21, 2023, Kharkiv, Ukraine ORCID: 0000-0003-1745-7026 (M. Kubiavka); 0000-0001-7689-239X (S. Lienkov); 0000-0002-9287-0817 (Y. Khlaponin); 0000-00025141-9886 (L. Kubiavka); 0000-0003-3221-2860 (A. Bernaz); 0009-0008-5940-1286 (K. Pyrogov)
A characteristic feature of today is the intensively developing and implemented informatization processes in almost all spheres of human activity. They led to the formation of a new information infrastructure, which is associated with a new type of social relations and the use of information and communication technologies (ICT).
The modern stage of informatization of education is also associated with the wide introduction of methods and means of information and communication systems and technologies into the education system, the creation of a computer-oriented information and communication environment on this basis, and the filling of this environment with electronic scientific, educational and management information resources, with the provision of opportunities for the subjects of the educational process to access the resources of the environment, use its means and services when solving various tasks. At the same time, without the use of modern means of processing natural language information, it is impossible to count on the convenience and effectiveness of using such systems in the conditions of quality management of education. It is clear that in such a complex field as the management of the quality of education, where the activities of managers and specialists are related to the processing of a significant amount of data, it is impossible to do without the use of natural language information processing tools. But the modern
2023 Copyright for this paper by its authors. Ukrainian market of information technologies (IT) does not offer educators a wide selection of instrumental intelligent systems that could be used with the lowest costs for processing natural language texts in the management of the quality of education.
Therefore, there is a need to create problem-oriented intellectual systems capable of "selecting" information necessary for managers and education specialists from the natural language presentation of various categories of projects in education and applying it to solve the problems of planning educational activities, managing the quality of education, managing the cost of education, managing risks, etc.
It is proposed to take the theory of information (non-force) interaction as a scientific and methodological basis for solving this task. Within the framework of this theory, not only the apparatus for forecasting the development of systems based on the interaction of people was created, but also the mathematical apparatus for the formalization of the soft component, also generated by human activity, was provided.
The purpose of the work is to create, using the mathematical apparatus of the theory of non-force interaction, models, methods and means of formalizing natural language information, on the basis of which it would be possible to create economically profitable, effective, simple and competitive ICT systems for the management of the quality of education and educational management. This work is dedicated to the development of such a reflexive intellectual system.
Education management information systems (EMIS) enable officials to be constantly informed
when developing the foundations of state education policy and to make final decisions based on reliable
information. Considering this, the requirements for the formation of educational databases are very
high, because poor-quality educational information can have far-reaching negative consequences,
including the adoption of incorrect management decisions regarding the distribution of limited
resources. In view of the above, the development of data information systems and their integration
should be the basis of state management of the education system [
The effectiveness of the construction of means of automated management of the quality of education
largely depends on the linguistic means of describing objects and management processes. In connection
with this, there are many studies on the construction of natural language processing systems [
The modern Ukrainian market of information technologies (IT) does not offer educators a wide selection of instrumental intellectual systems that could be used with the lowest costs for processing natural language texts in the management of the quality of education.
Known domestic and foreign natural language information processing systems are, for the most part,
experimental and highly specialized and do not fully satisfy the requirements of automated information
processing systems in education management. And although the Ministry of Education and Science of
Ukraine adopted the "Automated Information Complex of Educational Management" software and
hardware complex in 2021, its main purpose remains to ensure the transition to electronic document
management and access to aggregated depersonalized data for automatic reporting. In addition, the costs
of creating, maintaining, and modernizing such a complex are significant [
But nevertheless, without such means, the productivity of information and communication systems
when filling the information base of education management is significantly reduced, and the costs of
performing routine actions, such as the formal presentation and input of information, searching in the
information base, etc., increase [
The use of natural language processing systems in education management is associated with a number of specific features of this branch of activity, significant duration and complexity of the process of creating an educational product, significant changes in documentation during the implementation of such projects, incompleteness of documentation, low reliability of information sources, diversity of source documentation, which is formed at the request of various users, by the difficult formalization of most of the source documents. Accordingly, the requirements and limitations for natural language text processing systems are determined by mobility, reliability, high productivity, ease of changing the information base, and decentralization of the data processing process.
To date, within the framework of the theory of informational (non-force) interaction, not only the
apparatus for forecasting the development of systems based on human interaction has been created, but
also the mathematical apparatus for the formalization of the soft component, also generated by human
activity, has been provided. Today, in addition to managing the development of complex systems, the
theory is also used to manage enterprises [
The effectiveness of the construction of means of automated management of the quality of education
largely depends on the linguistic means of describing objects and management processes [
The resolution of this contradiction is related to: 1. With the division of incoming information flows into those that are formalized (and, accordingly, require traditional dialog software tools for their processing) and those that are difficult to formalize (require significant costs for formalization);
2. Using the results obtained in the theory of non-force interaction for the construction of natural
language information processing tools [
In order to build effective information and communication systems and technologies for the management of the quality of education, according to the authors, it is necessary to carry out the following preparatory work: • to study the object and situation in which the processing of natural language information is carried out; • choose a strategy, tactics, and means of processing natural language information; • determine the most influential forms of information presentation and develop models and methods of their formation with the help of all available means of presenting (proving) information; • to develop a system for analyzing the effectiveness of information provided and its correction in case of insufficient action.
The requirements for these tools are based on an understanding of the conditions in which educators find themselves, the sources and quality of the information received, as well as ways of forming various information representations.
Based on this, the following requirements are: 1. Simplicity. 2. Multifacetedness. 3. Algorithmic clarity. 4. The possibility of systematization of the main processes of information management. 5. Openness.
These requirements will be provided in further research: 1. By developing templates and methods for automatically determining the most influential ways of presenting information (simplicity).
2. The use of various information management tools, the choice of which will depend on the information needs regarding the required result (multifacetedness).
3. Formal methods of presenting management tools (algorithmic clarity).
model: parameters (1): encountered (2):
Development of standards for information interaction, focused on transforming the presented information into awareness (the possibility of automating the main processes of information processing).
The possibility of access to means of providing information from many workplaces (openness).
So, let's formalize the task of applying natural language information processing systems for decisionmaking in education quality management. This task is formulated as follows: it is necessary to perform such actions (or to make certain decisions) in the educational project that corresponds to the semantic and pragmatic content of the natural language text.
To solve this problem, and based on the informational redundancy of natural language, we will divide the input text into clusters, among which there will usually be those that determine the reaction required under the given conditions. Since it is impossible to know in advance which clusters these are, we will form a number of combinatorial regularities of the input text and, in the process of learning, select those that have the greatest impact on the decision. This is determined based on the following The set of elements of natural language affecting decisions is divided into clusters containing
= { }, = ̅1̅̅,̅̅, where is the input text; – is the cluster influencing the decision; is the number of clusters.
According to the frequency characteristics, it is possible to determine the approximate values of the probability of each of the decisions (each of the actions), provided that one or another cluster is ∈ , ∈ ∶ ( j / ) ≈ ( / ), where n(Rj / xi) – is the frequency of action (decision) Rj, provided that the cluster xi was present in the input text; p(Rj / xi) – is the probability of action (decision) Rj, provided that the cluster xi was present. 3.
For each action (decision) Rj given the unconditional probability p(Rj) and individual conditional probabilities p(Rj / xi),...,p(Rj / x1),...,p(Rj / xq), it is necessary to estimate the joint conditional probability of this solution (action) p(Rj / x1,…, xi,…, xq).
Since, in this interpretation, the decision-making process can be represented as the process of determining the most significant informational influences of the input text on a set of actions (decisions), we will use an approach based on the mathematical apparatus of the theory of non-force interaction to solve this problem. The measure of assessing the situation, relative to the solution in the theory of non-force interaction, is determination and awareness. The relationship between certainty and probability is given by formula (3): where p – decision probability; d – certainty of this decision.
The relationship between awareness and probability is given by formula (4): = { 0,5×√1− −0,5×√1−
+ 1− −2, ℎ ≥0,5 + 1− −2, ℎ <0,5
, =
0,5 √ ×(1− )
. 0 → ^ 0 →
⇒ 0 → ,
A measure of non-force action ∈ , = ̅1̅̅,̅̅ should reflect the difference in certainty and awareness of "before" and "after" changes in the subject area. And it is precisely the change in certainty and awareness that is the cause of the change in probability (5) where p0 – probability of decision (action) R0; d0 – certainty of decision (action) R0; i0 – awareness of the decision (action) R0; dj – certainty of decision (action) R0, which is formed by a cluster of input text ∈ , ij – awareness that corresponds to certainty dj; pj – probability of decision (action) R0 given by all clusters ∈ . (1) (2) (3) (4) (5)
At the same time, applying the mathematical apparatus of the Theory of non-force interaction, we obtain (6,7): case there is a cluster in the text ∈ to the certainty of decision-making R0 given by all clusters of the input text (8) ∀ , : 0 → ∑ ^ 0 → ∑ ⇒ ∑ = 0,5 + ∑ 2∑ , where ∑ – certainty of decision R0, which is formed by all clusters ∈ ; ∑ - awareness corresponding to certainty ∑ ; ∑ – the probability of the decision R0 under the influence of all clusters included in the input text В.
To solve this problem, we will apply the mathematical apparatus of the theory of non-force interaction. The concept of building reflex intellectual systems, which is based on the mathematical apparatus of the theory of non-force interaction, implements the principles of universal modeling of regularities in a given subject area, which are statistically manifested. The basis of the functioning of reflex intellectual systems is the method of developing an adequate reaction of the intellectual system to all external influences. This method ensures the selection of the most probable reaction of the intelligent system to a set of input influences, with known probabilities of choosing a reaction to each input influence, as well as to some combinations of input influences.
The reflex method directly calculates the response of the intellectual system adequate to the input
influences (Fig. 1). The adequacy of the reaction was confirmed in previous experimental studies
conducted by authors and covered in the relevant sources [
Input text
Non-forceful influences
Classes of reactions (6) (7) (8) 1
2 9 3 5 6 7 4 8
R
R E A C T I O N S
Note: those virtual objects that reflect the commonality of action (synergistic effect) are marked with a dotted line
The method of calculating an adequate reaction to the meaning of the input text in the education management project has the following stages:
Calculation of the certainty of reactions in the project relative to the effects of text clusters. From (5)-(6): ( ) = ±0,5√
( ) 1− ( ) + 1− ( ) − 2,
( ) ( ) = √ 2( ) + 1, ( / ) = ±0,5√
( / ) 1− ( / ) + 1− ( / ) ( / )
− 2, ( / ) = √ 2( / ) + 1,
Calculation of the total impact on decision-making managers or specialists who perform where ∆ ∑ ( ) – changing the certainty of the decision (action) Ri. changing the certainty of the decision (action) if there is a cluster in the input text . managerial actions (14): where p(Ri) – unconditional probability p(Ri) choice of decision (action) Ri; d(Ri) – certainty in the project regarding the decision (action) Ri; i(Ri) – awareness in the project regarding the decision (action) Ri; p(Ri /Aj) – the conditional probability of choosing a decision (action) Ri if present in the input text of the cluster Aj; d(Ri / Aj) – certainty in the project regarding the decision (action) Ri, if present in the input text of the cluster Aj; i(Ri /Aj) – awareness in the project regarding the decision (action) Ri, if there is a cluster in the input text Aj.
Calculation of the additional probability that input text clusters bring to the intelligent system ∆ ( / ) = ( / ) × ( ) − ( ) × ( / ), where ∆ ( / ) – changing the certainty of the decision (action) Ri if there is a cluster in the input ∆ ∑ ( ) = ∑
∆ ( / ), ∆ ∑ ( ) = √∆ ∑2 ( ) + 1, (9) (10) (11) (12) (13) (14) (15) (16) (17) where ∆ ∑ ( ) – change of awareness in the project, relative to the decision (action) Ri.
Calculation of the new certainty of decisions (actions) in the management of the quality of education (16):
̅̅(̅̅̅̅̅) = ∆ ∑ ( ) × ( ) + ( ) × ∆ ∑ ( ), where ̅̅(̅̅̅̅̅) – new determination of the project management's attitude to the decision (action) Ri.
Selection of the most definite decision (action) Rk (17): : ∈ ^̅̅(̅̅̅̅̅̅) = (̅̅(̅̅̅̅̅)).
During the creation of reflex systems for processing natural language information using the given method, it is necessary to solve two main tasks: to highlight in the input text and formalize the clusters that have the greatest impact, as well as the reactions of the management of the management of the quality of education with the accumulation of statistical information about the connections between them; adapt the reflexive algorithm to the features of natural language.
On the basis of the proposed model, one of the options for the application of the reflex system of natural language information processing has been developed.
t x e t t u p n I d n a s i s ly s a er n t a su lia lc it d in te fo lce ck se o l B m tse s y n
o s i f t o ce a n re on A c
Education is extremely important for socioeconomic development and cultural enrichment of society, providing people with relevant knowledge and skills to improve skills, and the ability to work productively in the conditions of further global development. For Ukraine, the task of effectively providing and organizing the education of pupils/students in the transforming education system, forming a new information mentality of all interested parties, is relevant. Solving this task requires constant monitoring and assessment of the state of the education system, which is based on the collection, processing of information, and analysis of educational data necessary to ensure the adoption of well-founded management decisions. Such information support makes it possible to implement modern education management information systems (Education Management Information Systems (EMIS)). The constant increase in the amount of professional knowledge and the high degree of development of information and communication technologies actualizes the problem of developing high-quality software products with elements of artificial intelligence designed to meet the information needs of users. The work of any intellectual system involves the processing of the knowledge base, which is its basis and has the property of gradually accumulating new knowledge. Intelligent systems must accumulate information, generalize it, and learn from their mistakes, using the knowledge base. Since the basis of an intellectual system, which is created as an application of the theory of non-force interaction, is the processing of various actions with the development of a reaction adequate to these actions, such systems are reflex (by analogy with the behavior of objects of living nature). The concept of building this reflex intellectual system implements the principles of universal modeling of regularities in a given subject area, which are statistically manifested. The method of developing an intelligent system response adequate to all external influences ensures the selection of the most probable response of the intelligent system to a set of input influences, with known probabilities of choosing a reaction to each input influence, as well as to some combinations of input influences.
In the process of this study, it was shown that the mathematical apparatus of the theory of non-force
interaction is quite simple and convenient for building natural language text processing systems. The
used mathematical apparatus will be the basis of the system of discrete processing of the semantics of
texts in the management of the quality of education and educational management. In further work, it is
planned to develop the structure and algorithms of the specified system and conduct its experimental
verification in production conditions for the management of educational processes. The use of the
developed new ICT will affect the development of the formed ICT environment, which is dynamic and
open, which will allow to improve the process of management of educational institutions, improve and
maximally direct their activities to increase the quality of education. The proposed model differs from
the existing ones by identifying differences in the conditional probabilities of the appearance of message
addressees as the amount of informational influence of fragments of natural language text, which made
it possible to automatically determine the most likely addressees of this text. The developed technology
meets the requirements for simplicity, is informative and easy to understand, and is minimally
expensive. Its foundation is based on the models and methods of influence process management
developed by the authors [
The author expresses his main gratitude to Professor Yury Teslia, he, like no one else, knows what non-violent interaction is. This study is based on the fact that the result of non-violent interactions is the construction of a general attitude to reality. The author also expresses his gratitude to Professor Serhii Lienkov, who mentors and teaches. Thank you, my dear sister Lyubov Kubiavka, for your faith in me and all-round support. Special thanks to Kyrylo Pyrogov for the quick translation of the work. Friends - you are the best!
[16] M. Kubiavka, Models and methods of managing information support in the conditions of hybrid warfare: synopsis for Ph.D. thesis in Engineering Science: 05.13.06, Kyiv. Taras Shevchenko nat. uni. of Kyiv, 2017. [17] Zha. Hong, Design of Information Teaching Management System Integrating Association Rule Mining Algorithm, BigIoT-EDU 2022: Application of Big Data, Blockchain, and Internet of Things for Education Informatization, pp. 439–445. doi: 10.1007/978-3-031-23950-2_48. [18] S. Lenkov, M. Kubyavka, L. Kubiavka, Y. Lеnkov, V. Shevchuk, Reflex Intellectual text processing systems: Natural language text addressing. CEUR Workshor Proceedings, 2019, pp. 85-95. [19] G. Fantoni, E. Coli, F. Chiarello, R. Apreda, Dell’Orletta, F. and Pratelli, G. Text mining tool for translating terms of contract into technical specifications: Development and application in the railway sector, Computers in Industry, Volume 124, 2021, URL: https://doi.org/ 10.1016/j.compind.2020.103357. [20] S. Kumar, A. Kar, and P. Ilavarasan, Applications of text mining in services management: A systematic literature review, International Journal of Information Management Data Insights, Volume 1, Issue 1, 100008, URL: https://doi.org/10.1016/j.jjimei.2021.100008. [21] M. Köseoglu, F. Mehraliyev, M. Altin, and F. Okumus, Competitor intelligence and analysis (CIA) model and online reviews: integrating big data text mining with network analysis for strategic analysis, Tourism Review, Vol. 76 No. 3, 2021, pp. 529-552. [22] S. Birunda, R. Devi, A Review on Word Embedding Techniques for Text Classification. In: Raj, J.S., Iliyasu, A.M., Bestak, R., Baig, Z.A. (eds) Innovative Data Communication Technologies and Application. Lecture Notes on Data Engineering and Communications Technologies, vol 59.
Springer, Singapore, 2021. URL: https://doi.org/10.1007/978-981-15-9651-3_23. [23] E. Bassignana, B. Plank, What Do You Mean by Relation Extraction? A Survey on Datasets and Study on Scientific Relation Classification. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop, Dublin, Ireland, 2022, pp. 67–83. [24] T. Haddadi, Q. Haddadi, T. Mourabit, A. Allaoui, and M. Ahmed, Automatic analysis of the sustainability of public procurement based on Text Mining: The case of the Moroccan ICT markets, Cleaner and Responsible Consumption, Volume 3, 100037, 2021, URL: https://doi.org/10.1016/j.clrc.2021.100037. [25] B. Baidena, J. Nimako-Koduab, V.K. Anyanfulc, D. Oppong, Management Information Systems and its Impact on Productivity in Higher Education: A Case of Colleges of Education in Ghana (IJC) Volume 45, No1, 2023, pp. 136-150.