Order	Time	ID	Full information
1	13:00-13:15	IT1052	Research and Evaluation on the Labor-Innovation Integration Education Path Based on Knowledge Graph under the Context of Artificial Intelligence Author(s): Jiang Jin-Gang, Zhang Jia-Wei, Shen Tao, Wang Kai-Rui, Bao Yu-Dong, Sun Jian-Peng Presenter: Jin-gang Jiang, Harbin University of Science and Technology, China Abstract: In order to better integrate innovation and entrepreneurship education and labor education in depth and organically and realize the two-way shaping of labor sentiment and innovation and entrepreneurship sentiment, this paper analyzes and summarizes the important value of exploring the path of labor-innovation integration and the realistic dilemmas of labor-innovation integration and explores the path of labor-innovation integration based on the concept of “ideological and political leadership, diversified synergy, and course-project-competition parallel drive" combined with knowledge graphs constructed based on artificial intelligence (AI) architecture, and evaluates the proposed path of labor-creation integration based on the multilevel fuzzy theory. This path can help the innovative thinking of college students to be tested and developed in practical labor and promote students to transform the labor results into actual economic benefits, thus promoting the sustainable development of society and the economy.
2	13:15-13:30	IT1079	Research on the Construction of AI-Based Personalized Learning Paths and Curriculum Design Author(s): Yuan Sun, Bojia Chen, Changhua Li Presenter: Bojia Chen, Jianghan University, China Abstract: With the rapid advancement of artificial intelligence and data-driven technologies, the education sector is being transformed from traditional uniform teaching models to student-centered personalized learning approaches, and the development of personalized learning paths and curriculum design has been identified as a critical direction for promoting high-quality educational development. This paper systematically reviews the current applications of artificial intelligence and data-driven technologies in personalized learning, thoroughly analyzes the opportunities and core challenges in technology implementation, and focuses on the mechanisms of AI-driven personalized learning path design and data-supported curriculum setting methods. Research indicates that the synergistic application of artificial intelligence and data-driven technologies can effectively achieve "teaching determined by learning," providing technical support for personalized education; however, addressing issues such as data security, algorithm fairness, and teachers' technological adaptation capabilities is considered a core prerequisite for successful technology implementation. This study can provide theoretical references for advancing personalized education practices in the educational field, optimizing the integration of technology and teaching scenarios, and facilitating the digital and personalized transformation of education.
3	13:30-13:45	IT2159	The Construction of an AI-based chatbot to Provide Formative Feedback on Argumentative Writing Author(s): Jing Wen, Xiaoyu Wang Presenter: Jing Wen, University of Electronic Science and Technology of China Abstract: The study explores students’ perceptions of the efficacy of an AI-based chatbot “IELTS Writing Tutor” in providing formative feedback on argumentative essays to first-year engineering students at a China-UK collaborative undergraduate program. Built on the Smartutor platform, the chatbot delivers structured feedback aligned with IELTS Writing Task 2 criteria, model answers, and reflection prompts. 29 students engaged in the research and received feedback on the three writing responses they wrote. This study employed a mixed-methods approach. Quantitative data were collected via a 5-point Likert scale questionnaire measuring perceived usefulness and ease of use of the chatbot, followed by SPSS-based descriptive analysis. Qualitative data included open-ended feedback from students, which was analyzed using thematic analysis. The study found participants rated the chatbot highly on both usefulness and ease of use, with all mean scores exceeding 4.2. While its core strengths are recognized in promptly identifying errors and delivering learning-focused feedback, the findings indicate that future development must focus on two key areas: advancing personalized learning experience and diversifying modes of interaction.
4	13:45-14:00	IT3198	LAVATour: A Career Digital Twin Platform for AI-Driven Trajectory Simulation and  Consultation Author(s): Xin Xie, Huimin Zhang Presenter: Huimin Zhang, Sichuan Technology and Business University, China Abstract: In the rapidly evolving landscape of the artificial intelligence (AI) era, traditional career planning tools—often based on static psychometric assessments like RIASEC or MBTI—fail to account for dynamic external variables such as algorithmic displacement and economic volatility. This paper introduces LAVATour, a web-based intelligent career navigation platform designed to operationalize the LAVA framework (Lifecycle, Algorithmic Displacement, Vocational Entropy, Active Inference). Unlike conventional tools that provide a single "snapshot" of career fit, LAVATour utilizes a React-based architecture and D3.js visualization libraries to render a dynamic, 100-year career trajectory. The system integrates a "Vocational Entropy" simulation engine that visualizes skill decay over time and an "Active Inference" module powered by Large Language Models (LLMs) to generate personalized, entropy-reducing milestone recommendations. We describe the system architecture, the mathematical logic underlying the trajectory generation, and the prompt engineering strategies used for context-aware consultation. LAVATour represents a paradigm shift from static career matching to dynamic career digital twinning, providing students with a "flight simulator" to navigate the complexities of the future labor market.
5	14:00-14:15	IT3283	Blockchain Practical Teaching based on Dual-Platform Author(s): Ke Xu, Deyou Tang Presenter: Deyou Tang, South China University of Technology Abstract: The industrial development of blockchain technology demands high-quality talents with both theoretical foundations and practical engineering capabilities. To address the prominent challenges in current blockchain talent cultivation, such as over-reliance on a single blockchain platform, fragmented practical teaching, and scarce high-quality teaching resources, this paper designs and implements a dual-platform practical teaching system underpinned by in-depth university-industry collaboration. Centered on two mainstream consortium blockchain platforms (Hyperledger Fabric and FISCO BCOS), the system develops two core courses: “Blockchain Technology and Applications” and “Blockchain Practical Training”, both featuring layered and progressive practical content. Additionally, a “Blockchain Practical Teaching Case Repository” covering finance, supply chain, public services, and other related fields is constructed. Teaching practice demonstrates that this system significantly improves student learning outcomes, strengthens the demonstrative impact of teaching reform, and gains wide recognition from the industry. It thereby provides a replicable model for cultivating blockchain engineering talents in higher education.
6	14:15-14:30	IT3341	Design and Feasibility Study of Dual-Graph Framework for Personalized Learning Based on Artificial Intelligence Author(s): Liguo Qu, Yunqi Hu, Mingxing Fang, Xiang Wang Presenter: Yunqi Hu, Anhui Normal University, China Abstract: Current adaptive learning systems inadequately characterize the dynamic evolution of knowledge structures and learner capabilities, making it difficult to form interpretable personalized diagnosis and path recommendations. Addressing the problem of long-term separation between knowledge graphs and capability graphs with lack of formalized linkage mechanisms, this paper proposes a Dual-Graph Framework intelligent tutoring system. Through Graph-regularized Deep Knowledge Tracing (G-DKT), the knowledge graph adjacency matrix is embedded into the loss function as a structural constraint to guide student capability assessment via knowledge topology. Simultaneously, a heuristic-rule-based navigation engine is designed to perform interpretable path planning on the knowledge graph based on capability graph states, forming a closed loop of "structure-constrained assessment and assessment-driven navigation", providing a technical blueprint with both pedagogical rationality and computational feasibility for personalized learning systems. A feasibility study using a photovoltaic maximum power point tracking (MPPT) teaching case suggests that the proposed framework can diagnose knowledge gaps and generate pedagogically coherent learning sequences by reconstructing learner capability states and producing topology-grounded, interpretable personalized paths.
7	14:30-14:45	IT3367	Application Research of Collaborative Multi-Agent Systems in Industrial Software Talent Cultivation Author(s): Min Huang, Jiarui Ma, Sun Bo Presenter: Min Huang, South China University of Technology, China Abstract: To address the challenge of maintaining learning continuity while providing personalized instruction in large-scale industrial software education, this paper proposes a collaborative multi-agent pedagogical framework. The framework establishes a four-layer agent collaboration architecture—comprising instructional coordination, skill decomposition, context management, and evaluation feedback—and introduces a “three-layer, multi-attribute, and relational” skill ontology model to enable dynamic granularity decomposition of complex engineering tasks. By integrating a hierarchical memory system and a tool invocation framework, the system achieves deep coupling with industrial software APIs and establishes a real-time, closed-loop verification mechanism spanning from operational execution to automated diagnosis. Results from a quasi-experimental study demonstrate that the proposed framework outperforms traditional instructional modes in terms of learning gains, instructional efficiency, and knowledge retention rates, while effectively providing differentiated scaffolding support based on the learners’ initial proficiency levels. This research provides a systematic solution for industrial software talent cultivation, ranging from skill modeling to intelligent intervention, and offers significant practical reference value for the development of intelligent pedagogical platforms integrated with industry-education synergy.
8	14:45-15:00	IT3400	Enabling Adaptive Chinese Idiom Teaching with Prompt-Guided Text-to-Cypher Knowledge Graph Queries Author(s): Yi Liang Presenter: Yi Liang, Tianjin Normal University, China Abstract: Idiom teaching presents a significant challenge in international Chinese education. The core of idiom teaching lies in decoding the complex relationships among characters, semantics, and structures within idioms. A knowledge graph is an effective model for representing these relationships. However, traditional methods of manually constructing and querying such graphs are inefficient, unable to support dynamic, personalized teaching needs. This study proposes a novel approach that integrates Prompt Engineering with Text-to-Cypher technology to intelligently mine idiom knowledge graphs, thereby facilitating the design of adaptive instructional paths. We first constructed an idiom knowledge graph based on the "Chinese Proficiency Grading Standards for International Chinese Language Education". Then we designed a set of prompt templates and processes to translate teachers’ natural language questions into efficient Cypher query sentences. This facilitates the retrieval of multi-dimensional relationships—such as character co-occurrence, semantic associations, and grammatical structures—directly from the Neo4j graph database. This study demonstrates that this approach can rapidly generate complex queries for teaching scenarios, thereby making explicit the underlying associative networks among idioms. These insights enable educators to design targeted learning paths and content based on individual learner profiles. This study provides a practical, data-driven framework for adaptive instructional design in Chinese idiom teaching, presenting a concrete application of AI-assisted curriculum development through the integration of prompt engineering and knowledge graphs.
9	15:00-15:15	IT4468	Smart Chinese: An AI-Native Platform for Academic Chinese Learning in Higher Education Author(s): Jian Wu, Weiying Chen, Teng Yao Presenter: Wu Jian, Zhejiang University, China Abstract: The rapid development of generative artificial intelligence has created new opportunities for transforming language learning systems from tool-based applications into integrated learning infrastructures. However, most existing Chinese language learning platforms primarily target general communicative competence and provide limited support for academic language development in higher education. This paper presents Smart Chinese, an AI-native platform that supports academic Chinese learning through a closed-loop architecture integrating standards-aligned assessment, adaptive learning, and disciplinary content generation. Large language models are employed to produce personalized multimodal learning resources, including conversational podcasts and interactive flashcards. To ensure reliability and compliance with national proficiency standards, the system adopts prompt-constrained generation and a dual-blind expert validation workflow. The implementation demonstrates how learner profiling, level-based class allocation, discipline-aware resource recommendation, and competency-based progression can be orchestrated within a unified system. By embedding academic content and human–AI collaboration into the learning pipeline, the platform bridges the gap between general language learning and academic literacy. This study provides a system-level perspective on AI integration in language education and illustrates how AI-native architectures can support scalable, standards-aligned, and context-aware learning in higher education.