In-depth Learning Research Articles

Quantitative evaluation of attraction intensity of highway landscape visual elements based on dynamic perception

In order to quantify the aesthetic attraction of visual elements of highway landscape space, the concept of visual attraction intensity of highway landscape space is proposed, and the factors and indicators of visual attraction elements are clarified. Based on the perspective of human factors, taking passengers as the main attraction, taking 30 landscape elements from three typical highway landscape scenes, namely mountain highway landscape, plain highway landscape and elevated highway landscape, as the research object. The real-time evaluation and data processing system of subjective landscape preference has been independently developed, and combined with the processing and analysis of human-machine-environment synchronous quantitative analysis technology and in-depth learning, a “subjective+objective” quantitative evaluation model of highway landscape visual element attraction intensity based on dynamic perception has been constructed, revealing the quantitative relationship between highway landscape visual attraction elements and attraction intensity, The quantitative method and action law of the attraction intensity of highway landscape and the dynamic aesthetic perception of the attraction subject is clarified. It is found that the change of highway planning alignment, high green view rate and low enclosure have positive effects on the subjective preference score of the landscape (0.690), and there is a significant positive correlation between green view rate and enclosure; the richer the landscape elements such as mountains, trees and water bodies, the stronger the visual attraction, but the complexity of the landscape elements increases the cognitive load of the area of interest. Among the physiological indicators, there are significant correlations between EEG and ECG (0.790), EDA and ECG (0.632), and subjective preference and passengers' eye-movement behavior and EDA; as well as the change pattern of attraction intensity based on the dynamic perception of tunnel area. The results show that the attraction intensity of visual elements of highway landscape space based on dynamic perception is consistent with human aesthetic preference, visual perception and physiological response, which verifies the scientificity, appropriateness and accuracy of eye-movement analysis method and physiological analysis method for quantitative evaluation of aesthetic preference of highway landscape.

Comparative study of deep learning models on the images of biopsy specimens for diagnosis of lung cancer treatment

ObjectiveTo compare the application value of different traditional deep learning models in diagnosing and classifying lung cancer. MethodsAccording to the biopsy samples of our hospital from January 2018 to November 2022, 37 patients treated in this department were selected as the study subjects. Nonsmall and small cell lung cancer biopsy specimens were obtained and stained. Two experienced pathologists diagnosed the biopsy specimens. Multiple in-depth learning methods were used to distinguish between cancer and noncancer biopsies. In this study, we compared the application value of traditional deep learning models in lung cancer diagnosis and Classification. ResultsThe study tested several popular CNN architectures based on image block classification: AlexNet, VGG, ResNet, and SqueezeNet, comparing two types of training schemes: training from scratch and fine-tuning the entire pretrained network. The AUC of the deep learning model is more reasonable (0. 8808–0. 9121). Except for Resnet-50, the AUC of the training is higher than that of the fine-tuning of the whole network. ConclusionDeep learning analysis can accelerate the detection speed of whole section images (WSI) and maintain a similar detection rate with pathologists.

LUCC Simulation Based on RF-CNN-LSTM-CA Model with High-Quality Seed Selection Iterative Algorithm

Land use/cover change (LUCC) models are essential for studying the profound impact of land use/cover dynamics on various aspects of the natural and social environment. Cellular Automata (CA) is widely used in the dynamic modeling of complex LUCC systems. In the traditional machine learning CA model, when using statistical methods to obtain neighborhood features, there is usually the problem that the spatio-temporal feature learning of neighborhood factors is insufficient. At the same time, the CA dynamic iteration module using the random seed selection mechanism often has the problem that the seed selection efficiency is very low. In this paper, taking the Chongqing Metropolitan Area as an example, convolutional neural networks (CNN)-Long Short-Term Memory Network (LSTM) is introduced to improve the learning effect of the traditional random forest (RF)-CA model in the spatial and temporal characteristics of neighborhood factors. CNN is used to extract the spatial dimension features of LUCC in the neighborhood, and the LSTM model is used to extract the time dimension features and long-term dependencies. At the same time, a high-quality seed selection iterative algorithm (HQSSIA) is used to improve the accuracy of the multi-land-use dynamic change model and the efficiency of the iterative algorithm. The results show that, the proposed model performs better than other models in simulating the LUCC from 2015 to 2020 (Kappa = 0.9684, FOM = 0.1744, Accuracy = 0.9829, F1 = 0.9641, Hamming = 0.0171) and from 2010 to 2020 (Kappa = 0.9599, FOM = 0.4662, Accuracy = 0.9785, F1 = 0.8113, Hamming = 0.0214). After introducing the CNN-LSTM model, the Figure of Merit (FOM) increased by 1.56% and 18.88% for 2015–2020 and 2010–2020. Compared with the CA model-based random seed selection algorithm, the FOM of the model using HQSSIA in the dynamic iteration module are improved by 11.60% and 24.79% for 2015–2020 and 2010–2020, and the operation efficiency of the dynamic iteration module is improved by about 19 times. Compared with the current mainstream LUCC models PLUS and FLUS, the proposed model has improved 14.38%, 37.55%, and 14.93%, 37.74% in FOM, respectively, for 2015–2020 and 2010–2020. The research shows that: (1) RF-CNN-LSTM-CA model not only retains the interpretability advantage of the traditional RF-CA model, but also improves the accuracy of the whole model by improving the spatio-temporal characteristics of neighborhood factors through in-depth learning; (2) the HQSSIA can quickly and accurately search for cells to be converted with higher conversion probability in the observed data, which can not only significantly reduce the time complexity of the model, but also improve the accuracy of LUCC simulation.

Financial Forecasting Method for Generative Adversarial Networks based on Multi-model Fusion

<p> To some extent, stock prices can reflect national economic development and residents’ living standards. However, current stock forecasts are mainly analyzed using the stock’s own price. It is found that the exchange rate and technical indicators are closely related to the fluctuations of stocks and stock indexes. In this study, a generative adversarial network (GAN) financial forecasting method based on multi-model fusion is proposed, which introduces the exchange rate and technical indexes of stock and stock index into the input data, and combines the characteristics of its own highest and lowest price, the closing price of stock or stock index is predicted by using the generated antagonistic network. The methods are as follows: firstly, the exchange rate features are de-noised by wavelet transform technology, the overall trend of exchange rate features is extracted, and the dimensionality of technical indicators is reduced by Principal Component Analysis (PCA). Then, a convolutional neural network (CNN) is used in the generator to extract the local features of the input data, and the attention mechanism is used to improve the prediction effect of the model. Finally, the prediction was made by long short-term memory network (LSTM). Convolutional neural network and multilayer fully connected neural network are used in the discriminant, and gelu is used as activation function in the hidden layer. This study selects the data of Zhuhai Port (000507.SZ), Xiamen International Trade (600755.SH), Orient Venture (600278.SH) and Shanghai Stock Exchange Index (SSE, 000001.SH) in the export industry from January 1, 2013 to December 31, 2019 as the experimental data set. The exchange rate price between RMB and US dollar at the corresponding time is selected as the exchange rate feature. Compared with the best in-depth learning financial prediction comparison model, the results show that the generation confrontation network proposed in this paper has improved by 3% to 15% in each experimental data set.</p> <p> </p>

Nesting the SIRV model with NAR, LSTM and statistical methods to fit and predict COVID-19 epidemic trend in Africa

ObjectiveCompared with other regions in the world, the transmission characteristics of the COVID-19 epidemic in Africa are more obvious, has a unique transmission mode in this region; At the same time, the data related to the COVID-19 epidemic in Africa is characterized by low data quality and incomplete data coverage, which makes the prediction method of COVID-19 epidemic suitable for other regions unable to achieve good results in Africa. In order to solve the above problems, this paper proposes a prediction method that nests the in-depth learning method in the mechanism model. From the experimental results, it can better solve the above problems and better adapt to the transmission characteristics of the COVID-19 epidemic in African countries.MethodsBased on the SIRV model, the COVID-19 transmission rate and trend from September 2021 to January 2022 of the top 15 African countries (South Africa, Morocco, Tunisia, Libya, Egypt, Ethiopia, Kenya, Zambia, Algeria, Botswana, Nigeria, Zimbabwe, Mozambique, Uganda, and Ghana) in the accumulative number of COVID-19 confirmed cases was fitted by using the data from Worldometer. Non-autoregressive (NAR), Long-short term memory (LSTM), Autoregressive integrated moving average (ARIMA) models, Gaussian and polynomial functions were used to predict the transmission rate β in the next 7, 14, and 21 days. Then, the predicted transmission rate βs were substituted into the SIRV model to predict the number of the COVID-19 active cases. The error analysis was conducted using root-mean-square error (RMSE) and mean absolute percentage error (MAPE).ResultsThe fitting curves of the 7, 14, and 21 days were consistent with and higher than the original curves of daily active cases (DAC). The MAPE between the fitted and original 7-day DAC was only 1.15% and increased with the longer of predict days. Both the predicted β and DAC of the next 7, 14, and 21 days by NAR and LSTM nested models were closer to the real ones than other three ones. The minimum RMSEs for the predicted number of COVID-19 active cases in the next 7, 14, and 21 days were 12,974, 14,152, and 12,211 people, respectively when the order of magnitude for was 106, with the minimum MAPE being 1.79%, 1.97%, and 1.64%, respectively.ConclusionNesting the SIRV model with NAR, LSTM, ARIMA methods etc. through functionalizing β respectively could obtain more accurate fitting and predicting results than these models/methods alone for the number of confirmed COVID-19 cases in Africa in which nesting with NAR had the highest accuracy for the 14-day and 21-day predictions. The nested model was of high significance for early understanding of the COVID-19 disease burden and preparedness for the response.

Transitioning to virtual reality learning in 5E learning model: pedagogical practices for science learning

ABSTRACT Virtual Reality (VR) has been widely adopted in educational research; however, its implementation in the classroom is still not as well known as in the research field. In other words, there is a gap between researchers and practitioners regarding the pedagogical issue. This study aimed to explore how VR could be applied in science classrooms using the 5E learning model approach, thus providing an overview of VR implementation from teachers’ perspectives. The method was a qualitative study in which it was validated using methodological triangulation, including open-ended questions, video-recording observations, and field notes. The finding revealed that learning science using VR for visualizing abstract concepts was suitable to be integrated with the 5E learning model, which provided an authentic opportunity for students to become curious and engage in the inquiry process. However, teachers should guide students to avoid misconceptions during the learning process. Moreover, this finding provides practical insights where an elaboration prompt is required before using a VR headset to turn the potential of VR into actual in-depth learning. Finally, 5E learning implementation can be integrated with other learning approaches, such as situated learning and guided inquiry learning, to suit the contents and teachers’ pedagogical knowledge.

Enhancing students’ understanding and performance in a distance-learning setting: evidence from an audit simulation at a GCC university

PurposeDrawing on experiential learning theory (ELT), this study aims to examine students’ performance and perceptions after performing an experiential learning activity (ELA) by completing a mini-audit simulation (AS) on the purchase and cash disbursement processes in a distance-learning environment at a Gulf Cooperation Council (GCC) university.Design/methodology/approachAdopting a mixed-methods approach, we collected quantitative and qualitative data from 176 students using the grade centre on Blackboard and their responses to a semi-structured questionnaire.FindingsThe pre-and post-simulation tests indicate significant improvement in students’ understanding and performance after performing the mini-AS. The students’ responses also provide robust evidence of student engagement, active participation and positive recognition of the AS’s value.Practical implicationsThis study has several implications: for the accounting education literature, how AS strengthens in-depth learning through the lens of ELT; for professional accounting bodies, informing the need to maximise the awareness and benefits of adopting simulations in accounting education and examination; and for educators, considering simulations in their ELAs to enhance student learning.Originality/valueThis study introduces a new authentic mini-AS instrument that can be adapted to a distance-learning setting, adds to the very limited studies in AS using ELT, uses a mixed-methods approach and explores students who learn in an Arabic-speaking country.

Multi-Window Identification of Landslide Hazards Based on InSAR Technology and Factors Predisposing to Disasters

Identification of potential landslide hazards is of great significance for disaster prevention and control. CNN (Convolutional Neural Networks), RNN (Recurrent Neural Networks) and many other deep learning methods have been used to identify landslide hazards. However, most samples are made with a fixed window size, which affects recognition accuracy to some extent. This paper presents a multi-window hidden danger identification CNN method according to the scale of the landslide in the experimental area. Firstly, the hidden danger area is preliminarily screened by InSAR deformation processing technology. Secondly, based on topography, geology, hydrology and human activities, a total of 15 disaster-prone factors are used to create factor datasets for in-depth learning. According to the general scale of the landslide, models with four window sizes of 48 × 48, 32 × 32, 16 × 16 and 8 × 8 are trained, respectively, and several window models with better recognition effect and suitable for the scale of landslide in the experimental area are selected for the accurate identification of landslide hazards. The results show that, among the four windows, 16 × 16 and 8 × 8 windows have the best model recognition effect. Then, according to the scale of the landslide, these optimal windows are pertinently selected, and the precision, recall rate and F-measure of the multi-window deep learning model are improved (82.86%, 78.75%, 80.75%). The research results prove that the multi-window identification method of landslide hazards combining InSAR technology and factors predisposing to disasters is effective, which can play an important role in regional disaster identification and enhance the scientific and technological support ability of geological disaster prevention and mitigation.

Product Recommendation System Based on Deep Learning

Abstract With the development of Internet big data and e-commerce, the widespread popularity of information, information acquisition and personalized recommendation technologies have attracted extensive attention. The core value of personalized recommendation is to provide more accurate content and services around users. The recommended scenarios are not uniform, and different dimensions need to be considered. For example, we are facing enterprises or individuals, different age groups, different levels of education, social life and other aspects. In this paper, the classic DNN (Deep Neural Networks) double tower recommendation algorithm in the recommendation algorithm is used as the ranking algorithm of the recommendation system. It is divided into user and item for embedding respectively. The network model is built using tensorflow. The data processed by the initial data through feature engineering is sent into the model for training, and the trained DNN double tower model is obtained. Recall adopts collaborative filtering algorithm, and applies tfidf, w2v, etc. to process feature engineering, so as to better improve the accuracy of the system and balance the EE problem of the recommendation system. The recommendation module of this system is divided into data cleaning as a whole. Feature engineering includes the establishment of user portraits, the analysis of multiple recall and sorting algorithms, the adoption of multiple recall mode, and the implementation of a classic recommendation system with in-depth learning. This makes the recommendation system better balance the interests of both the platform and users, and achieve a win-win situation.

Study on Characteristics and Constituent Elements of Learning Plan from the Perspective of Deep Learning

The in-depth learning from the perspective of learning science, which emerged in the 1970s, focuses on the social character and key ability of learners in the era of knowledge economy to adapt to and lead the development of modern society and become people who are beneficial to themselves and society. As a technical tool or spiritual value carrier of classroom teaching, learning plan attaches importance to "situational" complex learning research, pays attention to the questioning of cognitive subject "human value", shifts from attaching importance to the capacity of classroom knowledge to enhancing the depth of classroom knowledge, and aims to pursue an effective learning process of understanding and innovative learning of knowledge with important concepts as the core, and cultivate students' knowledge, ability and emotion necessary for in-depth learning. The study plan from the perspective of deep learning has the characteristics of knowledge questioning, problem thinking, hierarchical thinking and hierarchical gradient. It is composed of learning objectives, knowledge structure, problem design, scenario creation, reflection and evaluation, and forms a deep learning logical thinking framework with problem orientation as the guide and ideological development as the purpose throughout the process.

Detection and Tracking of Vehicles Using License Plate Recognition

The density of vehicles on roads has increased manifold over the last few decades. A seamless system to detect and track a particular vehicle can solve many problems, like traffic congestion, etc. This paper proposes the use of real-time data taken from closed-circuit televisions to detect and track the movements of vehicles. The feature extraction and classification are completely done by the process of faster RCNN (regional convolutional neural networks). The core work is based on region proposal networks. CNN furthers the generation of features; classification is done separately, but it is aided by RCNN, which includes deep learning as well as training the neural network. This is used along with the Kanade Lukas Tomasi algorithm to tack the desired features extracted. This method is simulated on MATLAB software.

Recognizing traffic signs with synthetic data and deep learning

Recently, in-depth learning about computer vision and object classification tasks has surpassed other machine learning (ML) algorithms. This algorithm, alike similar ML algorithms, requires a dataset for training. In most real cases, developing an appropriate dataset is expensive and time-consuming. Also, in some situations, providing the dataset is unsafe or even impossible. In this paper, we proposed a novel framework for traffic sign recognition using synthetic data and deep learning. The main feature of the proposed method is its independence from the real-life dataset, which leads to high accuracy in the real test dataset. Creating one-by-one synthetic data is more labor-intensive and costlier than providing real data. To tackle the issue, the proposed framework uses a procedural method, which gives the possibility to develop countless high-quality data that are close enough to the real data. Due to its procedural nature, this framework can be easily edited and tuned.

Research on Default Prediction Model of Corporate Credit Risk Based on Big Data Analysis Algorithm

In recent years, with the rise of many technologies such as big data and artificial intelligence, the digitalization and information transformation of enterprises have gradually penetrated into the financial industry. Based on different big data analysis algorithms, we aim to establish a default prediction model for corporate credit risk, further optimize different models, compare the model performance, and analyze the robustness of the optimal model. We collect 21 items of financial and non-financial index data from more than 1,000 listed companies, standardize, balance and normalize the data, use correlation coefficient to screen the index, and establish two in-depth learning models, convolutional neural network model and recurrent neural network model, based on Pytorch framework of Spark platform, to complete the model optimization, and compare them with two traditional machine learning models: random forest model and logistic regression model. Finally, the comparison experiment shows that the recurrent neural network is the optimal model with an accuracy rate of 0.93, a recall rate of 0.96 and a F1 value of 0.93. For the optimal recurrent neural network model, the robustness of the model is analyzed by modifying the number of indicators, changing the number of samples and eliminating non-financial factors. The results show that the evaluation indicators of the model do not change much under the three conditions, and the model has a good robustness.

Remaining Useful Life Prediction of Bearings Based on Convolution Attention Mechanism and Temporal Convolution Network

The prediction of the remaining useful life (RUL) of bearings is of great significance for reducing cost and increasing efficiency of mechanical equipment and ensuring healthy operation. Most of the traditional RUL prediction methods based on deep learning only analyze single sensor data, which makes the prediction accuracy fluctuate greatly and the reliability is low, and the contributions of data collected by different sensors to RUL prediction are inconsistent, which makes the low utilization of data lead to poor prediction results. To solve the above problems, a new RUL prediction framework with convolutional attention mechanism and temporal convolutional network (CAMTCN) is proposed. Firstly, the vibration signals collected by multi-sensor are normalized and processed directly as the input of the network. Secondly, an Efficient Adaptive Shrinkage (EAS) model is designed to adaptively eliminate noise interference and improve the accuracy of RUL prediction; Finally, we use the attention mechanism to design the convolutional attention subnet and combine it with the TCN to obtain the contribution degree of bearing degradation features collected by different sensors. Adaptive weight allocation captures more important degradation features to achieve end-to-end RUL prediction. Experiments were conducted on IEEE PHM 2012 dataset and XJTU-SY dataset respectively, and compared with some advanced in-depth learning methods. The experimental results show that this method can predict RUL more accurately on different bearing data sets, and has good generalization.

Дидактичний потенціал інтерактивних методів навчання на уроках “Я досліджую світ” у початковій школі

The article is devoted to the didactic potential of interactive teaching methods at the lessons «I Explore the World» in primary school, which consists in ensuring the active cognitive activity of students, the development of their creative thinking and the ability to cooperate. The historical aspects of the study of the concept of «interactive pedagogy», which was first introduced in 1975 by the German scientist Hans Fritz, are analyzed. The concept of «interactive», which in modern interpretation is understood as the ability for active interaction and dialogue, was studied. The article describes how didactic potential determines the possibility of pedagogical methods and tools to achieve maximum results in the development and education of students. The use of interactive methods promotes in-depth learning of knowledge through practical activities, discussion and exchange of ideas. The authors provide examples from educational practice of how students have the opportunity to actively interact with the material, express their opinions, analyze and summarize information. Forms and methods of educational technologies of interactive learning, which are recommended to be used in the process of teaching the course «I Explore the World» have been analyzed. The criteria for formulating students' achievements in the process of interactive learning at the lesson «I Explore the World» are outlined. And the peculiarities of the use of interactive tasks at individual stages of the lesson are also analyzed. Thus, interactive methods contribute to the development of communication skills, social competence and independence of education seekers, creating a favorable climate for learning and active knowledge of the surrounding world. The authors proved that the use of interactive methods (in particular, «unfinished sentence», «associative bush», «brainstorming», «aquarium», «microphone» and others) allows elementary school students to systematize and generalize knowledge, develop creativity and creative thinking, promotes involving each student in active discussion and analysis of important issues, develops their ability to cooperate and express their opinions in a reasoned manner. The further development of this direction involves the research of new interactive methods and their adaptation to the needs of primary school students, which will contribute to the enrichment of the teaching methodology and the improvement of the quality of education. Keywords: primary school, educational process, «I explore the world» lessons, knowledge acquisition, interactive learning methods, potential, didactic potential.

Stacked Ensemble-Based Type-2 Diabetes Prediction Using Machine Learning Techniques

Diabetes is a long-term disease caused by the human body's inability to make enough insulin or to use it properly. This is one of the curses of the present world. Although it is not very severe in the initial stage, over time, it takes a deadly shape and gradually affects a variety of human organs, such as the heart, kidney, liver, eyes, and brain, leading to death. Many researchers focus on the machine and in-depth learning strategies to efficiently predict diabetes based on numerous risk variables such as insulin, BMI, and glucose in this healthcare issue. We proposed a robust approach based on the stacked ensemble method for predicting diabetes using several machine learning (ML) methods. The stacked ensemble comprises two models: the base model and the meta-model. Base models use a variety of models of ML, such as Support Vector Machine (SVM), K Nearest Neighbor (KNN), Naïve Bayes (NB), and Random Forest (RF), which make different assumptions about predictions, and meta-models make final predictions using Logistic Regression from predictive outputs from base models. To assess the efficiency of the proposed model, we have considered the PIMA Indian Diabetes Dataset (PIMA-IDD). We used linear and stratified sampling to ensure dataset consistency and K-fold cross-validation to prevent model overfitting. Experiments revealed that the proposed stacked ensemble model outperformed the model specified in the base classifier as well as the comprehensive methods, with an accuracy of 94.17%.

A Study on the Evaluation Model of In-depth Learning for Oral English Learning in Online Education

A Study on the Evaluation Model of In-depth Learning for Oral English Learning in Online Education

The construction of curriculum learning pattern of master of education: theoretical foundation and practical framework

The implementation and operation of the “Curriculum for Education” is an important medium and carrier for nurturing the in-depth learning literacy of postgraduates, especially the implementation and operation of the principle courses provide an important platform and guarantee for cultivating the principle thinking of M.Ed. Combined with the analysis of research data, this paper finds that (1) the learning pattern of M.Ed. courses has the characteristics and practical characteristics of openness, initiative and self-adaptability. (2) The practical framework for constructing the learning pattern of the M.Ed. programme consists of two modules that lay out progressive learning tasks that point to “academic competence in the curriculum” and “practical competence in teaching”. (3) Guiding M.Ed. students to participate in course evaluation and promoting the formation of M.Ed. learning experiences are the core mechanisms of M.Ed. course learning patterns in practice.

Investigating adult learners' experiences from using slow reading as a pedagogical approach

This article explores the utilization of slow reading as a pedagogical tool aimed at enriching the learning experience in higher education, with a particular emphasis on fostering meaningful engagement with academic texts. Drawing upon insights from a qualitative case study conducted within a Norwegian vocational teacher training program, this research challenges the conventional belief that effective learning is synonymous only with speed and efficiency. By adopting a qualitative approach inspired by critical pedagogy, this study investigates the multifaceted dimensions of slow reading as an in-depth learning strategy. The inquiry includes a diverse range of data sources derived from a research and development project, encompassing the perspectives and reflections of both students and educators. The study's findings shed light on how individuals perceive and engage with slow reading as a pedagogical approach and its impact on their learning experiences. Notably, the study reveals that slow reading serves as a catalyst for fostering interaction, contemplation, and the exchange of experiences and novel ideas, ultimately contributing to the creation of an inclusive and dynamic learning environment. In conclusion, this research underscores the potential benefits of integrating slow reading as one among several study techniques within higher education.

Strengthening Student Engagement in Learning Through Use of Digital Tools

This paper analyzes teaching and learning in the context of the implementation of digital technology in educational practices. On the basis of the data obtained from focus group discussions and creative workshops with primary school teachers, ways are identified to recognize and strengthen student engagement in learning through the use of digital tools, based on artificial intelligence and integrating learning analytics. As evidenced by the analysis of the research data, students’ cognitive, emotional, and behavioral engagement in learning increases due to the use of digital learning tools. To support and strengthen student engagement, teachers need to rethink their habitual educational practices, anticipate the barriers to engagement caused by digital technologies and the ways of coping with them, make use of digital tools for personalized and in-depth learning, and apply effective means to manage classrooms and interactions between students.