With the advancement of economic globalization and web-based educational technologies, traditional static assessments fail to meet practical demands of international business English teaching and learning due to scenario disconnection and lack of real-time, personalized feedback. This study develops an artificial intelligence-driven, web-based dynamic assessment system with a cloud-edge collaborative, four-layer architecture. It collects multimodal data from simulated business scenarios, uses domain-adaptive Transformer models for language analysis, and applies reinforcement learning to adjust assessment parameters dynamically. Experiments show that system scores correlate highly with expert ratings (r = 0.91), with an average module accuracy of 91.8%. The system effectively tracks learners' ability development and narrows individual differences. The findings innovate business English assessment from summative evaluation to formative diagnosis, supporting the optimization of web-based business English learning and teaching.
This study designed a mobile web-based adaptive microlearning system for solfeggio education, grounded in cognitive load theory. A dual closed-loop framework integrating real-time pitch detection via web audio application-programming interface, behavioral logging, and dynamic difficulty adjustment was designed to reduce extraneous load and activate germane load. A 4-week quasi-experiment with 120 preservice music teachers demonstrated significant performance gains in the microlearning group compared to face-to-face instruction, with baseline gaps narrowing from 24 to 9 points. Mediation analysis confirmed that performance improvement was driven by the path of extraneous load reduction → germane load increase, accounting for 35% of the total effect. The system offers a replicable paradigm for skill-based mobile learning.
Through a four-week quasi-experiment, this study explores the influence of virtual reality (VR) immersion teaching on college students' intercultural communicative competence (ICC). The study divided 120 sophomores into a VR group and a traditional teaching group. The results showed that the total score of ICC in the VR group increased, which was significantly higher than that in the control group, especially in cultural awareness and emotional input. Based on the theory of experiential learning and embodied cognition, a model accounting for immersion, cultural empathy, and ICC shows that immersion plays a core intermediary role in the development of ICC; moreover, behavioral data (about such elements as head rotation and voice duration) are moderately correlated with ICC promotion. The reliability of the scale is high (α = 0.93). This paper puts forward a three-step teaching process consisting of situation creation, immediate feedback, and reflection internalization, which provides a practical paradigm for cross-cultural teaching in colleges and universities.
In this study, the authors investigate an innovative digital practice model for youth ideals and beliefs education, based on a data-driven “perception–judgment–action–feedback” closed-loop system. Through a quasi-experimental design involving 1,264 students, the authors compare the effects of traditional, basic digital, and advanced closed-loop models. Results show that the advanced digital model significantly outperforms the others, with higher task completion rates (92%), deeper interaction frequency, and greater identity improvement. The model also extends user retention and effectively supports marginalized learners. Findings demonstrate that dynamic, timely feedback and smart clustering strategies are crucial for enhancing educational effectiveness in the digital age. The study provides empirical evidence and practical pathways for transforming ideals and beliefs in education into a more responsive, participatory, and adaptive process.
Web-based and blended learning environments present a challenge to teachers in choosing appropriate instruction strategies to address the needs of various learners. This paper presents the Decision Support Framework of Web-Based Instruction (DSF-WBI) to support teachers in blended English courses by enabling them to make intelligent and interpretable decisions. DSF-WBI combines fuzzy reasoning dedicated to pedagogical transparency with machine learning trained for predictive accuracy and tested on four datasets representing learner engagement, sequence-level language practice, writing proficiency, and analytic evidence of English learner proficiency. The framework had an average accuracy of 91.9, cross-domain strength (less than 4% reduction in performance), and a 95% adoption rate among teachers. DSF-WBI is an effective tool that combines interpretability and performance, enabling educators to make transparent and data-informed instructional choices that support explainable AI applications in web learning.
Generative AI offers new opportunities to realize truly personalized, web-based language instruction. This study designed a generative-AI-powered system that dynamically generated and adaptively delivered English learning resources tailored to individual learners. By integrating bidirectional-encoder-representations-from-transformers-based semantic embeddings with structured learner attributes, the system built fine-grained user profiles to guide prompt-engineered content generation. A real-time difficulty adjustment module modulated resource complexity based on interaction data, while a graph neural network–Markov decision process model optimized personalized learning paths. Experimental results showed significant improvements in content relevance, task completion, learning gains, and user satisfaction compared to static approaches, demonstrating the pedagogical value of generative AI in responsive web-based learning environments.
In the era of deep penetration of intelligent technology, ideological and political courses in colleges and universities are facing the dual challenges of declining participation and lagging evaluation. To solve these problems, based on a comprehensive teaching pilot in a university, this paper constructs a double closed-loop teaching framework of data circulation and control circulation and introduces three core technologies, namely edge computing, emotion recognition, and adaptive push, to conduct millisecond-level perception and dynamic intervention on classroom state. In this paper, an artificial intelligence (AI)-empowered ideological and political education model of low invasion, high morality, and strong ethics has been successfully verified through experiments. AI can be used not only as a tool to improve teaching efficiency but also to promote the formation and solidification of students' internal values through accurate real-time intervention, providing a replicable and scalable technical path for the future development of wisdom education.
This study proposes a dual-loop AI model to enhance college students’ competency development through systematic industry alignment. The first loop leverages modular skill tracking, long short-term memory–attention networks, and Shapley additive explanations to deliver real-time, personalized feedback during learning. The second loop integrates enterprise job role profiles to align training with labor market demands. Deployed across multiple institutions in China, the system served 500 students and achieved sub-120ms latency with transparent visualizations. Results show the intervention group improved by 18.7 percentage points in employment competency, significantly exceeding the control group’s 5.9-point gain (p < .01). Key factors include feedback timeliness and job role alignment. With a high user satisfaction rating of 4.3/5, the model enables a closed-loop talent pipeline between academia and industry. These findings advance intelligent, data-driven approaches to competency-based education and workforce readiness.
Traditional piano fingering instruction often suffers from low efficiency and limited personalization. To address these limitations, the authors present a web-enabled AI system that supports real-time, adaptive fingering guidance in educational settings. The system fuses Musical Instrument Digital Interface (MIDI) data and three-dimensional finger joint tracking to model both musical and ergonomic dimensions, and it employs a hybrid graph neural network and reinforcement learning framework to generate comfortable, accurate fingering suggestions. A confidence-based difficulty adjustment strategy and miniature haptic feedback create a closed-loop learning experience, while cloud–edge deployment ensures responsiveness and privacy. Evaluated through technical performance, electromyography fatigue, cognitive load, and motivation metrics, the system significantly improves students’ consistency, engagement, and satisfaction—without increasing teacher workload. This work offers a scalable, AI-driven approach to skill-based music education.
The authors proposed a web-based AI framework that dynamically generates personalized learning paths for vocational learners. By integrating deep knowledge tracing, semantic path encoding, and multi-objective reinforcement learning, the system balanced completion efficiency, satisfaction, and dropout risk in real time. A/B tests across three vocational domains showed 8–14% gains in completion and 0.7–1.2-point boosts in satisfaction over baseline. The approach is deployable, interpretable, and adaptable to resource-constrained settings.
The integration of intelligent e-learning systems (IES) has transformed undergraduate education in Chinese higher education institutions, enhancing learning effectiveness (LEF) through technology-driven pedagogy. This study examines the impact of IES adoption on undergraduate LEF, with learning engagement, psychological motivation, and cognitive load optimization as mediators. Using structural equation modeling, data from 349 university students in China were analyzed to assess these relationships. The findings reveal that IES adoption significantly enhances LEF, with engagement and cognitive load optimization playing mediating roles, while psychological motivation shows no direct effect. The study contributes to educational technology research and underscores the need for pedagogical alignment and structured digital learning strategies to maximize IES benefits in higher education.
In response to the disconnect between language and music teaching in preschool education, this study focuses on the value of integrating digital storytelling with music into early childhood education. Preschool children aged three to six were randomly divided into an experimental group and a control group. A mixed research method was used over a 12-week intervention period. The experimental group implemented age-based digital storytelling with music-integrated teaching, supported by home-school collaborative activities, while the control group adopted traditional separated teaching. Research has shown that digital storytelling with music, through multisensory stimulation and situational interaction, aligns with the developmental characteristics of young children. Age-specific instructional design and home-school collaboration mechanisms can effectively promote the coordinated development of language and music abilities. These results provide a scientifically feasible practical paradigm to integrated language and music education in for preschool settings.
Based on the context of international curriculum reform and core literacy, this study explores regional differences in college teachers’ educational habits and professional qualities across ecological zones and proposes a theoretical model of ecoregional influence. Through literature review and content analysis, core elements are identified; the model is developed via inductive and deductive reasoning, and research instruments are designed. Empirical data were collected using a self-developed questionnaire and analyzed with structural equation modeling in Amos. Findings indicate that the ecoregional environment significantly shapes teachers’ educational habits and quality, revealing pronounced regional disparities. These results contribute to advancing teacher literacy development theory, informing ecological curriculum implementation, and supporting teachers’ professional growth and academic advancement.
This study proposes a deep learning-driven model to personalize instruction in university vocal music courses within a web-based learning environment. By integrating neural feedback mechanisms, the system dynamically monitors student performance and cognitive states in real time, generating adaptive learning pathways tailored to individual needs. It provides learners with immediate, data-driven feedback and equips instructors with comprehensive, actionable insights for timely pedagogical adjustments. An immersive, scenario-based component further enhances engagement and practical skill development. Empirical results demonstrate that the model significantly improves both teaching quality assessment and student learning outcomes, offering a robust, data-informed framework for personalized music education.
This study introduces a dual-loop intelligent feedback system designed for web-based English writing instruction. The system’s state-aware mechanism dynamically integrates automated scoring with teacher intervention. A semester-long quasi-experiment involving 100 undergraduates demonstrated its effectiveness: average feedback latency was reduced by half, and student engagement in proactive revisions increased substantially. Multimodal data analysis showed that the intervention not only improved surface-level writing features but also prompted a structural shift in writing competency. Students achieved balanced progress in higher-order dimensions such as content and organization with teacher-initiated feedback acting as a catalyst for competency leaps, particularly among lower-proficiency learners. These findings indicate that intelligently orchestrated, data-informed feedback pathways can effectively reconcile instructional scalability with deep, personalized writing development in digital learning environments.
In the era of digital transformation, the Internet of Things (IoT) technology is increasingly influencing educational paradigms, particularly in campus culture construction and aesthetic-political education. This study explores how IoT technology reshapes campus ecosystems by enhancing interactivity, personalization, and intelligence. Through a comprehensive longitudinal case study at a Chinese university from 2019 to 2023, the authors examined the integration of IoT into various educational activities, including traditional lectures, art exhibitions, school history events, environmental themes, and innovation activities. The findings indicated that IoT applications significantly boosted student engagement and ideological-political literacy, with notable improvements in innovation awareness (32.3% increase). Additionally, the study identified three dimensions for deep integration—goal fusion, content fusion, and technical fusion—and proposed practical pathways for other institutions aiming to integrate digital transformation into education.
VR hardware “demonstration viewing” lacks closed-loop feedback and system evaluation in art teaching. In this paper, a three-dimensional immersion teaching framework of VR-STE based on “perception-interaction-feedback-re-creation” is proposed, and 90 students are divided into two groups through the dynamic adaptation of double-cycle mechanism. The VR group completes six periods of creation in immersion environment, and the control group adopts traditional screen teaching. The results show that the spatial understanding of VR group is 1.8 points higher, and the creative quality is improved by 12.4%, which is better in “material application” and “composition integrity”. VR-STE constructs a three-dimensional framework of VR-STE through the double-cycle structure, multi-source fusion evaluation and the demonstration of VR's promotion of multi-dimensional creative ability, which compensates for traditional teaching shortcomings and offers a replicable 3D art teaching paradigm, with reference value for higher education reform, industry training, and XR education development.
Under China’s new curriculum reform, English teaching evaluation in secondary education calls for process-oriented and data-driven solutions. It still struggles with inefficient extraction of heterogeneous grade data and weak prediction of teaching effectiveness. This article presents an intelligent framework integrating a hidden Markov model and a deep neural network (DNN). The hidden Markov model module uses text block segmentation and maximum likelihood estimation to parse unstructured web-based grade data with high precision. The structured data are then used to train the DNN for effectiveness prediction. This framework achieved over 95% extraction accuracy on 1,200 teacher-course records. The DNN model outperformed multiple linear regression with R2 = 0.89 and mean absolute error = 1.73. This scalable solution supports real-time teaching quality monitoring and teacher development, offering a technical route for web-based, process-oriented English teaching evaluation.
With the growing use of technology in higher education, particularly in ideological and political education (IPE), effective sentiment analysis of student feedback is essential. Current methods often overlook the complexity of IPE discourse, leading to limited insights. This paper presents the value-aware sentiment analysis model, which combines sentiment lexicons, attention mechanisms, and a multi-task transformer architecture to analyze feedback across emotional polarity, value alignment, and cognitive dissonance dimensions. A real-time visualization dashboard allows instructors to adjust teaching strategies based on student sentiments. Experimental results confirm the model’s effectiveness, offering a scalable solution for enhancing IPE through artificial intelligence feedback analysis.
This study presents the design and evaluation of an immersive Japanese language learning system based on multimodal artificial intelligence to address limitations of traditional learning environments, including minimal interaction and delayed feedback. The system integrates speech, visual, and emotion-recognition technologies to enhance learner engagement through context-aware interactions and adaptive content delivery. The study examines key multimodal fusion techniques, identifies limitations in existing models, and optimizes algorithm design and interaction patterns accordingly. Empirical results indicate that the proposed system significantly improves oral accuracy, learning motivation, listening and speaking skills, and overall learning engagement, confirming its effectiveness and scalability. This work provides a practical framework for the development of immersive language learning systems and offers insights for future research on personalized and interactive artificial intelligence–supported education.
