Research on the Construction of Intelligent Supported Teaching Mode of College English Classroom in the Context of New Liberal Arts
Published Online: Mar 19, 2025
Received: Oct 18, 2024
Accepted: Feb 11, 2025
DOI: https://doi.org/10.2478/amns-2025-0393
Keywords
© 2025 Dandan Shi, published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
As the concept of new liberal arts education is rising, much attention has been paid to the application of digital technology in the field of education and teaching. In this context, it is essential for college English teachers to build an intelligent-supported college English teaching mode, use digital tools to create an effective English learning environment, improve the quality and efficiency of college English teaching, combine the needs of new liberal arts education with the goal of educating people, and cultivate students’ innovative ability and interdisciplinary literacy [1–3].
The advent of artificial intelligence brings a variety of opportunities and challenges for college English teaching. Along with the further development and application of AI technology, college English teaching will present a more diversified and personalized development trend, bringing students a more personalized and efficient learning experience [4–7]. Artificial intelligence technology can not only simulate a real language environment, so that students can learn English in real English communication scenarios, but also provide personalized learning resources and suggestions according to students’ interests and needs [8–11]. However, AI also has limitations, such as the inability to fully understand human emotions and cultural backgrounds, which may become an obstacle to developing education [12–13].
Regardless of technological advances, the role of the teacher is irreplaceable. Teachers can give students emotional support and encouragement and guide them to establish correct values [14]. In the English classrooms of universities, the innovation and change of English teaching methods based on artificial intelligence technology can enable teachers to more effectively grasp the learning status of students, so as to provide targeted tutoring and assistance, and deepen the communication and interaction between teachers and students [15–17]. Therefore, in college English teaching, artificial intelligence technology should be combined with the advantages of teachers to jointly promote the overall development of students [18–19]. Only in this way can we truly realize the goal of education and cultivate English talents with international vision and cross-cultural communication ability.
With the background of new liberal arts, this paper constructs an intelligent teaching system and proposes an intelligent interactive teaching method with the integration of artificial intelligence technology and college English classroom as the orientation and the improvement of college English teaching quality and effect as the starting point. Data mining, machine learning, intelligent search, personalized recommendation, and other technologies are used to design the AI system module for English teaching. Using statistical analysis, questionnaire and other research methods, we analyze the changes in the performance of English listening, speaking, reading and other abilities before and after teaching, so as to explore the application and effect of the intelligent interactive mode in English teaching.
Driven by the wave of digitalization and intellectualization, new technologies, new business forms and new modes are emerging and changing rapidly. The humanities and social sciences, as important tools for understanding and transforming the world, urgently need to undergo a profound reform from the traditional mode to the new liberal arts to meet the needs of the times. The “Declaration of New Liberal Arts” was released at the “New Liberal Arts” Working Conference organized by the Ministry of Education, which marks the beginning of a brand-new stage of development of the “New Liberal Arts”, and also signals that we are about to see a profound change in the field of liberal arts education. The “New Liberal Arts” is committed to enhancing the effectiveness of liberal arts education, aiming to break through the barriers between traditional disciplines, subvert the inherent disciplinary patterns, and actively explore innovative ways of cultivating talents in liberal arts, with a view to opening up a new situation in liberal arts education, as well as injecting new vitality into the cultivation of talents and paradigms of thinking in other disciplines.
The English classroom needs an interactive teaching environment because the learning environment in a university is not the same as that in an English classroom. Many universities have more conditions and opportunities to create an interactive English environment through English corners, foreign teachers, exchange students, internships in foreign companies and many other channels, but there is no such interactive environment inside or outside the classroom. In the traditional classroom language environment, it is very difficult for students to understand English, let alone have a smooth and comfortable English conversation, which is the root of the problem of “deaf English”. However, artificial intelligence can provide technical support for the interactive English teaching environment. Through artificial intelligence technology, computers are able to comprehensively process text, graphics, images, sound and other media information, establish a logical connection between this information and integrate an intelligent system, which is the most important feature of the intelligent system is that it can provide a variety of interactive methods of English teaching. The English teaching interactions provided include voice, text, and graphical interface interaction, and these three interactions can be performed simultaneously to achieve the corresponding teaching effects. The teaching interactions provided are shown in Figure 1.
Teaching interaction
Through the literature analysis method, we found that although the relevant research and development of teaching systems based on the market is in full swing, their relevance and practicality are comparative low. A variety of disciplines are mixed, and there is no access to artificial intelligence software specifically for teaching English. Moreover, many teaching practices show that English teaching urgently needs the support of technology to optimize the teaching method, and the characteristics of technology can precisely make up for the insufficiency of the English teaching process, especially in the simulation of the natural language environment and the creation of an English learning atmosphere, which is the biggest advantage of technology.
Therefore, to optimize students’ learning outcomes and provide them with the immersive learning experience, this paper starts from the perspective of modern educational technology students, takes the integration of information technology and curriculum as the orientation, applies the relevant theories of pedagogy, new curriculum theory and other disciplines and the research methods of literature analysis method and interview method, explores the application scheme of AI in English teaching, and focuses on the use of AI such as natural language comprehension, machine learning and intelligent searching technologies, attempting to carry out the design of AI-based English teaching system modules, all of which will be used as a substitute later in this paper. The system designed in this paper has four application modules, namely, auxiliary teaching module, knowledge explanation module, exercise recommendation module and environment simulation module, and the relationship between the main application modules is shown in Figure 2. With the help of the figure, this paper can further analyze the specific functions of the four modules in the system, namely, the auxiliary teaching module, the knowledge explanation module, the exercise recommendation module and the environment simulation module. First of all, the user body of the auxiliary teaching module is the English teacher, while the user body of the other modules is the student. In this module, English teachers can upload teaching videos through the system interface to organize teaching courseware. Teachers can use the data mining technology to assess the overall situation of the English test, including grading students’ English test, analyzing and evaluating students’ knowledge system and comprehensive language use ability, and providing teaching references.
Main module diagram
Secondly, students can apply what they have learned by previewing and reviewing knowledge points through video or text through the knowledge explanation module in the system. Thirdly, students can understand their own English knowledge system and test their own mastery or lack of basic knowledge through the recommended exercises module in the system, in which the types of exercises cover four aspects of English: reading, listening, writing, and conversation, aiming to comprehensively examine students’ listening, speaking, reading, and writing skills. Finally, after students complete the exercises, the module automatically generates the corresponding assessment levels and learning suggestions. In addition, teachers can also conduct online tests for students through this module to measure their real knowledge mastery. Due to the fact that the examination system is based on a large and scientific question bank, test question analysis and examination results analysis will be provided after the examination is completed. Therefore, these not only provide references for English teachers to teach according to their abilities, but also help learners to personalize their learning. The users of the environment simulation module are also students. Finally, special emphasis should be placed on the environment simulation module in the system, which takes the chatbot as the main part, where the user inputs voice or text information system through the interface to intelligently obtain the input information, and the use of text dialogues, video dialogues, role-playing dialogues, scenario dialogues and other functions can realistically simulate the actual scenarios of the English common life and create an active and authentic language learning atmosphere. The functions of each main module and sub-module are described in detail below.
Considering the dynamics of learners’ interests in the system, the recommendation module can provide learners with more accurate resource recommendation results. The category of exercises can directly show learners’ interest and preference for exercise resources, and the length of time learners study exercises reflects their interest and preference to a certain extent. Using exercise category information to analyze learners’ interests can improve the analysis of their exercise learning needs, thus providing them with more accurate personalized services and improving their satisfaction with the system.
The learner’s preference weight for a particular exercise category under the current time window is calculated with the formula:
Then the learners’ preference weights for each of the thirteen exercise categories under the current time window are calculated, and finally the dynamic interest model of learners based on exercise categories under the current time window is obtained:
Creating a Learner-Exercise Category Scoring Matrix Calculate the learner’s weights for the exercise categories in the same time window to obtain the learner-exercise category scoring matrix Calculating the similarity and find the set of similar users Based on the cosine similarity formula, we calculate the learner-learner similarity Finding High Frequency Resources Assuming that the set
The accuracy of word pronunciation, the phonetic intonation of pronunciation, and the number of long and short sentences used were used as criteria for listening and speaking assessment when learners used the intelligent teaching system to influence their English listening and speaking skills. Accurate English pronunciation is the first condition for students to learn English listening and speaking well. Students recite English and compare it with the standardized recitation version, and make targeted adjustments in voice and intonation according to the differences with the standardized version, which can not only improve students’ English pronunciation level, but also enhance their sense of English. The voice evaluation software should develop recording and playback functions, so that students can compare and imitate according to the voice intelligent evaluation module, listen and practice at any time, and play back according to their needs, so that students’ recitation learning need is free from the confinement of the classroom. In addition, the voice assessment software should improve the network sharing platform for teachers and students to use and interact with each other, so that students can record their recitation assignments and upload them to the learning platform, and teachers can comment on and tutor them according to the recordings, so as to better help students’ learning of listening and speaking in English. Moreover, the speech assessment software should develop the function of contextual dialogues, using vivid dialogues which are closely related to students’ lives, guiding students to speak English in the context, expressing what they think and need in English, and stimulating students’ interest in practicing listening and speaking in English.
The computational process of the K-means algorithm is to iterate repeatedly in order to achieve the purpose of minimizing the sum of squares (
In this paper, we characterized the online learning behavior data of 140 students majoring in English at a university. Teachers can see the data of students’ participation in various activities during the period through the platform data feedback after the class. This includes following, dictating, listening, reading, vocabulary, and other tasks assigned by the teacher. The data enables teachers to gauge the student’s performance in that task. At the same time, teachers and students themselves can also view the current period’s learning reports, students’ personal learning files and other aspects of data in their personal space as shown in Table 1. Through in-depth analysis of each learner’s personal information, including but not limited to name, student registration number, college and university, major subject, class, year of enrollment, etc., and then exhaustively sorted and summarized the amount of tasks completed, the percentage of tasks completed, the learning progress of the course videos, the completion of the tests of the chapters, the cumulative length of the videos watched, the number of times of participation in the discussion, and the frequency of the chapter review. Key online learning data aids teachers in targeting tutoring to students based on individual performance, and students can evaluate their own learning through data and progress with the assistance of team members and teachers. At the same time, students and teachers can also evaluate students based on their individual performance and related qualities to form a holistic assessment of the student.
The overall training data of the students in the background
| Name | Training completion number | And read times | Average score | Dictation | Hearing test | Audiometry | I heard the number of times |
|---|---|---|---|---|---|---|---|
| 1 | 85/86 | 75 | 92.2 | … | … | 8 | 4 |
| 2 | 85/86 | 76 | 91.0 | 4 | 96.6 | 9 | 5 |
| 3 | 84/86 | 70 | 97.3 | … | … | 8 | 4 |
| 4 | 83/86 | 79 | 91.5 | … | … | 9 | 5 |
| 5 | 80/86 | 67 | 82.7 | … | … | 9 | 5 |
| 6 | 80/86 | 68 | 90.3 | 4 | … | 10 | 4 |
| 7 | 82/86 | 67 | 92.0 | 4 | 91.0 | 2 | 3 |
| 8 | 81/86 | 65 | 91.3 | 1 | 97.2 | 7 | 3 |
| 9 | 79/86 | 70 | 86.2 | 2 | 9 | 2 | |
| 10 | 77/86 | 72 | 88.2 | 3 | 5.2 | 10 | 5 |
| …… |
The online learning behavior data was clustered using K-means method. And the distribution of clusters is shown in Figure 3. We can analyze:
Cluster distribution
Cluster 1 (gray dots) accounts for 25.2% of the total. Learners in this cluster show moderate learning engagement and learning achievement. Their characteristics include moderate frequency of course visits and assignment submission rates. Compared to Cluster 1, these learners may be experiencing some barriers and difficulty in the learning process, but they still have some ability to learn.
Cluster 2 (red dots) accounted for 63.5% of the total. Learners in this cluster exhibit high levels of academic engagement and higher academic achievement. Characteristics include a higher frequency of course visits, assignment submission rates, and test scores. It can be concluded that these learners have good study habits and strong learning abilities.
Cluster 3 (blue dots) accounts for 11.3% of the total. Learners in this cluster exhibit lower levels of learning engagement and academic performance. Characteristics include a lower frequency of course visits, assignment submission rates, and test scores. It can be concluded that these learners may have a lack of motivation or face additional obstacles to learning.
In the questionnaire on the use effect of the English exercise learning recommendation system, the survey analyzes the learners’ satisfaction with the system, mainly from three aspects: whether the learners support the behavioral data to be accessed and used, whether the system resources are designed to be comprehensive, and whether the recommended exercises are in line with the personalized needs. The results are shown in Figure 4, 56.4% of learners strongly support the use of learner behavioral data, 27% support the use of behavioral data, and most learners support the use of behavioral data to varying degrees to analyze learners’ personalized characteristics and recommend exercise resources for them. 60% of the learners think that the resources of the system are very comprehensive, and 23% think that the resources of the system are more comprehensive. On the whole, most learners think that the exercises in the system are comprehensive, and a few of them think that the exercises can be supplemented again. Combined with question 10 in the questionnaire on the effectiveness of the use of the English Exercise Recommendation System, regarding the shortcomings of the system, some learners have pointed out that regarding the design of resources for exercises, it is possible to consider adding English explanations to the exercises. Some learners pointed out that the design of resources for exercises could be considered to include English explanations of exercises, and that further improvement of the system could be considered by analyzing the feedback data from the learners. At the same time, 62.4% of the learners and 18% of the learners think that the exercise recommendation meets the demand to different degrees, which shows that most learners think that the exercise recommendation function of the system can basically meet the learners’ personalized exercise demand, indicating that the recommendation function of the system has a good recommendation effect. Overall, most learners’ satisfaction with the system is recognized to varying degrees, and the analysis of the learners’ satisfaction survey shows that the English Exercise Learning Recommendation System has a certain degree of effectiveness.
System satisfaction survey
In this study, the students’ questionnaires were designed from five competence dimensions, namely, language perception, language expression, language comprehension, information acquisition, and affective attitudes, respectively, and were used to consider the changes in students’ competence and learning attitudes. Among them, the examination of students’ English proficiency is mainly carried out based on the four proficiency dimensions of language perception, language expression, language comprehension, and information acquisition.
In the intelligent interactive teaching experiment carried out in this study, SPSS software was used to statistically organize the pre- and post-test questionnaire data of the students in the experimental class, and then a paired-sample t-test was done to obtain the test results as shown in Figure 5. Among them, the post-test mean of the language perception ability dimension increased by 0.32 compared with the pre-test mean, indicating that the language perception ability of the students in the experimental class was improved after the experiment compared with the pre-test.
Test class student ability matching sample statistics
In the dimension of language comprehension ability, the language comprehension ability of the students in the experimental class gained significant improvement after the experiment than before the experiment (Sig.=0.001). The post-test mean of the language expression ability dimension is 2.786, and the intelligent interactive teaching mode can improve students’ language ability. The post-test mean of students’ information acquisition ability dimension is 0.329 higher than the pre-test mean, and the intelligent interactive teaching mode can also improve information acquisition ability. This shows that after the teaching experiment, the abilities of the experimental class students in language perception, language expression, language comprehension, and information acquisition have been improved to different degrees. The significance values are all 0.001 < 0.01, indicating that before and after the experiment, there are significant changes in the experimental class students’ abilities in the four dimensions of language perception, language expression, language comprehension and information acquisition. This shows that this teaching experiment has significantly improved the English proficiency of the students in the experimental class.
Before the experiment, the students in the experimental class were pre-tested through the English listening and speaking test questions, and after the experiment, the test was conducted again, the content of which was mainly centered on the intonation, stress, legato, word accuracy and fluency of English pronunciation, and the scores were uniformly scored in accordance with the standards of the listening and speaking assessment, and the results of the test are shown in Fig. 6. It can be seen that the students in the experimental class had a significant improvement in the scoring grades of intonation, stress, legato, word accuracy and fluency of English listening after the experiment, and the post-test grades were all more than grade 6, which indicates that the overall level of English listening and speaking of the students in the experimental class had a significant improvement after the experiment.
Test results
By collating the pre-test scores and post-test scores of the experimental class, a paired-samples t-test was conducted to observe and analyze whether the students’ English reading ability changed before and after the experiment. The data are shown in Table 2. The average post-test score of the experimental class is 4.666 points higher than that of the pre-test, Sig.=0.001<0.01, indicating that there is a significant difference between the pre and post-test scores of the students in the experimental class, which means that the intelligent interactive teaching mode constructed in this study improves the students’ reading scores.
The individual sample group statistics of the periodic test results
| Test grades and final results of the experimental class | ||||||
|---|---|---|---|---|---|---|
| Mean value | Case number | Standard deviation | Standard error mean | T | Sig. | |
| Pretest | 12.359 | 50 | 8.184 | 1.157 | -14.353 | 0.001 |
| Aftertest | 17.025 | 50 | 6.734 | 0.952 | ||
| Periodic test | ||||||
|---|---|---|---|---|---|---|
| Class | Case number | Mean value | Standard deviation | Standard error mean | T | Sig. |
| 1 | 50 | 16.849 | 6.153 | 0.864 | -1.475 | 0.040 |
| 2 | 50 | 18.936 | 7.612 | 1.075 | ||
The teaching experiment in this study was divided into two phases, and the reading quiz was carried out after the completion of the first phase. After the quiz, results were statistically organized, SPSS software was used to carry out independent samples t-test, in order to analyze whether there was a significant difference in the reading achievement of the two classes after the first stage of the teaching experiment. In the stage test, the mean scores of the two classes were 16.849 and 18.936, respectively, and class 2 (experimental class) was 2.087 higher than class 1 (control class). According to the test data, the significance 0.04 < 0.05 indicates that there is a significant difference in the stage test scores of the two classes.
This paper builds an English teaching system using artificial intelligence-related technologies and proposes an intelligent interactive teaching mode. Through a questionnaire survey and teaching experiment, we will explore the practical application of intelligent interactive teaching mode. Understanding the students’ English learning under the intelligent interactive teaching mode, some meaningful conclusions were obtained:
Characterizing the learning data of English majors, we get 3 types of clustered distributions: low, medium and high levels of learning engagement and learning achievement. Most learners are basically satisfied with the exercise recommendation function of the intelligent teaching system. The Intelligent Interactive Teaching Mode significantly improves students’ language perception, language expression, language comprehension, information acquisition, and affective attitude abilities. The overall level of English listening and reading was significantly improved after the intelligent interactive teaching.