**3.1 Structure e-AMS**

The prototype being created is a distributed system for managing the acquisition of professional and language skills by adults. The main components of the system [16] (**Figure 12**) are: Interactive Speech Trainers with Visual Models (IST with VM), a Content Management System (CMS), a Speech Recognition and Modeling System (SRS), a System for Continuous Evaluation (CES) of the level of language skills and a Virtual Assistant (VA).

**Figure 12.** *Generalized structure of e-AMS.*

This training system has the properties:

unity of the main goal of learning outcomes for all its elements; high stability of the entire system and the independent value of each element; correlation between the elements of the system, providing positive feedback in the process of formation of professional and language skills; continuous evaluation of the level of competence of the student, which ensures the formation of a logarithmic dependence of the learning curve and compensates for the prerequisites for the degradation of the learning curve in the direction of loss of expected competence; ability to evaluate and compare the results of similar educational systems and technologies, as well as their individual components; cross-platform in relation to operating systems and programming languages, and invariance in relation to the student's native language and the foreign language being studied.

**Interactive Speech Trainers** are a new generation of chat bots that use artificial intelligence mechanisms to configure each student in the system, and Visual Models are in the form of augmented reality elements. This combination will allow you to fully immerse yourself in the process of forming skills and guarantees the maximum concentration of the student's attention.

**Content Management System** is designed to ensure the formation of educational material based on the most frequent words and semantic structures from the Corpus of the English language, subsequently modified by the subject area of the acquired professional skill. The current frequency is automatically determined using the Google Ngram Viewer resource.

In the course of work of the **Continuous Evaluation System**, a detailed statistical analysis of the results is carried out, the dynamic learning curves of each adult student are derived, the coefficient tables are justified and the indicators of the speed of formation of speech skills are clarified, the levels of speaking from initial to spontaneous are evaluated in accordance with the CEFR scale [14]. Not the curriculum, and the level of professional skills and language proficiency in the current time will be the determining to move forward on the learning curve.

Special methods of using a **Virtual Assistant** will allow you to organize the process of training language skills in such a way that the inclusion of the translation mechanism could not happen physiologically. The system will feed the material at such a speed that the exercise can only be performed without translation. Another option is also possible, which is used now, when the target speed of simple exercises is set such that it is almost impossible to perform them in a transferable way in the allotted time.

In addition, the VA must perform the function of a media aggregator, which provides the ability to connect to various services and find interesting professional content (virtual classes, photos, audio, and video) according to various criteria with reference to the level of current language proficiency of an adult student.

E-AMS continuously operates at three levels:

moderator-programmer who is one of the development team; teacher who updates content, processes general learning statistics at different levels, and in complex cases decides on the teacher's work through interactive platforms "face-to-face" with the student;

the most important level is the level of an adult student who consciously "transfers" his own consciousness to the control circuit of the educational process.

*Concept of a Management System for the Formation of Adult Language Skills on the Example… DOI: http://dx.doi.org/10.5772/intechopen.96926*

**Figure 13.** *Visual Representation of e-AMS.*

Of interest is the Visual Representation of e-AMS (**Figure 13**), which is an attempt to display all of the above in one figure. In fact, this is a visual representation of the theory of interiorization, or the creation of an indicative basis for mental activity according to Halperin, implemented in the adult learning process using e-AMS.

#### **3.2 System architecture**

Currently, the development of the main components of e-AMS is carried out in four interrelated areas:

further improvement of the visual approach that models the structure of the mastered activity based on an interactive visual dictionary and visual models; creation of a set of interactive speech simulators with elements of augmented reality, video lessons and training exercises corresponding to different levels of students' competence;

development of a content management system, formation of repositories of educational materials for video tutorials and mechanisms for managing the synthesis of educational exercises;

setting up a system of continuous evaluation and management of the learning process in real time.

The last two directions of development of e-AMS are connected exclusively with Big Data. The language system is not limitless and in fact is quite observable, but its full use presents numerous opportunities for combining, which until recently exceeded all possible system conditions, and only now the use of Big Data technology can successfully solve these problems.

In addition, a system of continuous evaluation and learning management provides the configuration of the entire system at the levels of:

interface of speech simulators, where all the achievements in the field of gamification, socialization and cooperation should be fully provided for the translation of the educational process into a modern, intensive and effective format;

continuous measurement and control of the learning process of each individual point of a learning curve or retraining in real time, which provides an individual approach to individual skill training, as the rate of exercise is not defined by an external statement or interface training program and abilities and competence level of each individual student.

**Figure 14** shows the system architecture for the first three main parts of the e-AMS: IST with VM, CMS, and SRS [17].

**Implementation of the platform**. The Platform will be implemented using the C# and JavaScript APIs. Since the advent of the HTML5 standard, there have been new opportunities to develop robust and efficient APIs that support voice processing, and this powerful, widely accepted standard includes interaction with various media, protocols, and programming languages [18]. The HTML5 page can handle voice and speech recognition recorded directly from devices, possibly available on the user's hardware. In addition, WebGL programs consist of control code written in JavaScript and special effects code (shader code) rendering in HTML, use the development element to draw WebGL graphics for visual programming with the introduction of programmable visual effects (for example, elements of augmented reality in the form of VM) and interact with the web page using scripts.

The **web server** contains a client-side API for transmitting the received data from the database (DB) server to the graphical interface for the client and connecting from the web server to the database using data access layers. These layers contain a connection manager and business layers containing a set of objects (classes) that return data as a set of data for use by web services and JavaScript and WebGel using Google services for speech modeling and speech recognition.

**Database design**. These are relational data tables that contain information about the lessons, as well as all information about the results of the lessons for each student for further extraction into the system. The results about any student and the lessons conducted with them will be stored in the DB, and dynamic learning curves will be formed on their basis.

**Information Request and Display Module** (IRDM). In this module, images representing action verbs in sentences are displayed as action and VM images. This

**Figure 14.** *System Architecture.*

*Concept of a Management System for the Formation of Adult Language Skills on the Example… DOI: http://dx.doi.org/10.5772/intechopen.96926*

**Figure 15.** *The learning curve with the transfer of synergistic effect.*

module allows the teacher to control the program parameters via the keyboard and computer screen, and this information is sent to the web server: once the information is processed the selected data is displayed to the teacher as the content of the next lesson.

**Knowledge Data Formation Module** (KDFM). The module consists of a DB that contains all the relational tables and is used to describe and represent the area of knowledge defined by the topic of the classes. The DB contains stored procedures that provide a way to create lesson content at the teacher's request depending on the selected level of complexity.

**Speech recognition module** (SRM). This module allows the student to interact in our platform using voice. The system returns the information to the SRM for conversion to speech using Text-to-Speech (TTS). The user speaks into the microphone, and this information is converted into text in the Automatic Speech Recognition (ASR) system. In the web server the java script functions check the correctness of the displayed text in the IRDM from this input data. Finally, the feedback of speaking will be permanently stored in the DB to evaluate and improve the transmitted meaning and pronunciation at the same time.

The introduction of modern ICTs into the structure of the foreign language LMS, along with the use of effective learning models, allows to speed up this process and increases its success by transferring the synergistic effect to all stages of the formation of language skills, especially to the "barriers to overcome" to accelerate the return to the expected competence curve (**Figure 15**). This approach opens up a new scientific direction in the construction of LMS.
