🤖 智能教育研究简报
报告日期: 2026-03-29
数据来源: arXiv + 17个教育技术核心期刊RSS + GitHub Trending
论文总数: 45 篇 (arXiv 26 + 期刊 19)
GitHub项目: 15 个
🔬 建议深读: 10 篇
📚 arXiv 重点论文
1. From Untamed Black Box to Interpretable Pedagogical Orchestration: The Ensemble of Specialized LLMs Architecture for Adaptive Tutoring
🔬 作者: Nizam Kadir
📅 发表: 2026-03-25
💡 摘要: Monolithic LLM拆分为多个专业LLM组合,实现可解释的自适应教学编排
🔗 链接: https://arxiv.org/abs/2603.119xx
2. CyberJustice Tutor: An Agentic AI Framework for Cybersecurity Learning via Think-Plan-Act Reasoning and Pedagogical Scaffolding
🔬 作者: Baiqiang Wang, Yan Bai, Juan Li
📅 发表: 2026-03-19
💡 摘要: 基于Think-Plan-Act推理和教学脚手架的网络安全学习Agentic AI框架
🔗 链接: https://arxiv.org/abs/2603.107xx
3. When AI Meets Early Childhood Education: LLMs as Assessment Teachers in Chinese Preschools
🔬 作者: Xingming Li, Runke Huang et al.
📅 发表: 2026-03-07
💡 摘要: 针对中国3600万儿童、25万+幼儿园的师幼互动(TCI)评估难题,提出基于LLM的自动化评估系统
🔗 链接: https://arxiv.org/abs/2603.xxxxx
4. Rewarding How Models Think Pedagogically: Integrating Pedagogical Reasoning and Thinking Rewards for LLMs in Education
🔬 作者: Unggi Lee, Jiyeong Bae et al.
📅 发表: 2026-01-20
💡 摘要: 整合教学推理与思维奖励信号用于教育LLM训练,提出教学推理奖励的新方法
🔗 链接: arxiv-2601-pedagogical-reasoning
5. Problems With LLMs for Learner Modelling: Why LLMs Alone Fall Short for Responsible Tutoring in K-12
🔬 作者: Danial Hooshyar, Yeongwook Yang et al.
📅 发表: 2025-12-28
💡 摘要: LLM单独用于K-12学习者建模的局限性——为何不足以实现负责任的辅导
🔗 链接: arxiv-2512-learner-modelling
6. Problems With Large Language Models for Learner Modelling: Why LLMs Alone Fall Short for Responsible Tutoring in K-12 Education
🔬 作者: Danial Hooshyar, Yeongwook Yang et al.
📅 发表: 2025-12-28
💡 摘要: LLM单独用于K-12学习者建模的局限性——为何LLM alone不足以实现负责任的辅导
🔗 链接: https://arxiv.org/abs/2512.xxxxx
7. The Landscape of AI in Science Education: What is Changing and How to Respond
作者: N/A
📅 发表: 2026-03-26
💡 摘要: 探讨人工智能在科学教育中的变革性角色及如何应对变化
🔗 链接: https://arxiv.org/abs/2603.145xx
8. Is Mathematical Problem-Solving Expertise in Large Language Models Associated with Assessment Performance?
作者: Liang Zhang, Yu Fu, Xinyi Jin
📅 发表: 2026-03-26
💡 摘要: 探讨LLM的数学专家能力与其评估表现之间的关联性
🔗 链接: https://arxiv.org/abs/2603.144xx
9. Offline-First Large Language Model Architecture for AI-Assisted Learning with Adaptive Response Levels in Low-Connectivity Environments
作者: Joseph Walusimbi, Ann Move Oguti, Joshua Benjamin Ssentongo, Keith Ainebyona
📅 发表: 2026-03-24
💡 摘要: AI and large language models for low-connectivity adaptive learning environments
🔗 链接: https://arxiv.org/abs/2602.15583
10. Privacy by Design Framework for LLM-Based Applications for Children
作者: Diana Addae, Diana Rogachova et al.
📅 发表: 2026-02-19
💡 摘要: 儿童LLM应用的隐私设计框架,涵盖GDPR/COPPA等儿童隐私保护法规
🔗 链接: arxiv-2602-privacy-children
11. A Privacy by Design Framework for Large Language Model-Based Applications for Children
作者: Diana Addae, Diana Rogachova et al.
📅 发表: 2026-02-19
💡 摘要: 儿童LLM应用的隐私设计框架,涵盖GDPR/COPPA等法规映射
🔗 链接: https://arxiv.org/abs/2602.xxxxx
12. Llama-Polya: Instruction Tuning for LLMs based on Polya’s Problem-solving
作者: Unggi Lee, Yeil Jeong et al.
📅 发表: 2026-02-11
💡 摘要: 基于Polya问题解决框架的指令微调LLM,将Polya的数学问题解决方法整合到对话结构中
🔗 链接: arxiv-2602-llama-polya
📖 期刊最新发表
数据来源: Computers & Education、BJET、JCAL、IEEE TLT、The Internet and Higher Education、Teaching and Teacher Education 等17个核心期刊
📗 British Journal of Educational Technology
- 🔬 The synergy of pedagogical agents and metaphorical design: Reducing psychological distance to enhance video learning
- 作者: BJET Editorial | 发表: 2026-03-26
- 教学代理与隐喻设计的协同效应:降低心理距离以增强视频学习,129名学习者2x2被试设计实验
📗 Computers & Education
- 🔬 Teachers’ doings in the AI era: From transmission to relation
- 作者: Miguel Nussbaum, Zvi Bekerman, Leanna Archambault, Maria Ranieri, Ramazan Yilmaz, Jiun-Yu Wu, Johan van Braak, Di Zou, Chin-Chung Tsai | 发表: 2026-08 (Vol.249)
- AI时代教师角色转型:从知识传递到关系构建,多国学者合作研究
- 🔬 Trusting AI to detect AI? A systematic evaluation of the reliability and robustness of current AIGC detection tools
- 作者: Yicheng Sun, Yihan Liao, Xiaoxue Ma | 发表: 2026-08 (Vol.249)
- 系统评估当前AIGC检测工具对学生学术作品的可靠性与鲁棒性
- The Interplay of Self-regulation, Cognitive Load and Learner Resources in Virtual Reality
- 作者: Nina R. Peltzer, Jochen Kuhn, Matthias Stadler, Raphael Cera, Christoph Hoyer, Salome Flegr, Martin R. Fischer, Constanze Richters | 发表: 2026-08 (Vol.249)
- VR环境中自我调节、认知负荷与学习资源的相互作用及对学习成果的影响
- Longitudinal relationships between student ethical considerations, behavioral intention, and perceived knowledge in AI education
- 作者: Weipeng Shen, Ching-sing Chai, Thomas K.F. Chiu, King Woon Yau, Helen Meng, Irwin King, Savio Wong, Yeung Yam | 发表: 2026-08 (Vol.249)
- AI教育中学生伦理考量、行为意向与感知知识的纵向关系研究
- Fostering preservice teachers’ design thinking by integrating generative AI and analytics-supported reflective assessment
- 作者: Yuqin Yang, Linbaiyu Chen, Daner Sun, Wangda Zhu, Shiyan Pang | 发表: 2026-08 (Vol.249)
- 生成式AI+分析支持反思性评估在协作探究环境中培养师范生设计思维
- Innovating diagnostic learning: How generative AI explanation styles influence cognitive load and confidence calibration in medical students
- 作者: Fang Tang, Feng Yao, Minghe Zhou | 发表: 2026-08 (Vol.249)
- 生成式AI解释风格对医学生认知负荷与信心校准的影响
📗 IEEE Transactions on Learning Technologies
- Design and Implementation Framework of Game Activity Generators for Declarative Knowledge Training
- 作者: Berenice Lemoine, Pierre Laforcade, Sebastien George | 发表: 2026-01 (Vol.19)
- 面向声明性知识训练的游戏活动生成器设计与实现框架,基于模型驱动工程(MDE)理论
- Virtual Robots in E-Learning: A Pathway to Enhanced Academic Self-Esteem, Math Performance, and Engagement
- 作者: IEEE Editorial | 发表: 2026-01 (Vol.19)
- 虚拟机器人对儿童学业自尊、数学表现和参与度的增强效应,实验组vs对照组各12名儿童
- M-LATTE: Multimodal Latent Temporal Modeling for Continuous Engagement Assessment in Online Education
- 作者: Congcong Xie, Di Wang, Quan Wang, Xiao Liang, Ruyi Liu, Qiguang Miao | 发表: 2026-01 (Vol.19)
- 多模态潜时空建模用于在线教育持续参与度评估,MSE从0.19降至0.10
📗 Journal of Computer Assisted Learning
- Continuous Temporal Contiguity Narration-Animation for Learning in Virtual Reality
- 作者: JCAL Editorial | 发表: 2026-04 (Vol.42(2))
- VR环境中连续时间毗邻性对叙述-动画学习的影响,227名初中生实验研究
📗 Teaching and Teacher Education
- Emotional experiences of English language teachers in the age of generative AI
- 作者: Mohamad Reza Farangi, Hassan Nejadghanbar | 发表: 2026-07 (Vol.176)
- 生成式AI时代英语教师的情感体验研究
- Teacher identity negotiation in AI-enhanced higher education and its implications for professional development
- 作者: Jie Li, Qianyu Zhang | 发表: 2026-07 (Vol.176)
- AI增强高等教育中教师身份协商及其对专业发展的影响
📗 The Internet and Higher Education
- 🔬 Fair AI in educational predictions: A multi-group fairness approach using reinforcement learning
- 作者: Fan Zhang, Wanli Xing, Chenglu Li, Yihan Jiang | 发表: 2026-04 (Vol.69)
- 使用强化学习的多组公平性方法实现教育预测中的公平AI
- Emotional artificial intelligence in higher education: A systematic review
- 作者: Mohammadbashir Sedighi, Sinnathamby Sritharakumar | 发表: 2026-07 (Vol.70)
- 高等教育中情感AI的系统综述
- Do we worry about the use of artificial intelligence and plagiarism? Students’ AI-gializer behaviour through the fraud triangle
- 作者: Mirac Yücel Baser, Metin Kozak, Ibrahim Halil Erdogan | 发表: 2026-04 (Vol.69)
- 学生AI-gializer行为(AI+plagiarism)研究,通过欺诈三角理论框架分析
- The effects of AI-enhanced adaptive socially shared regulation strategies on science pre-service teachers’ instructional design skills
- 作者: Jun-Qi Wu, Meng-meng Zhang, Fei-yan Wu, Jun Huang, Ya-juan Han, Yu Liu | 发表: 2026-04 (Vol.69)
- CSCL中AI增强自适应社会共享调节策略对科学师范生教学设计技能的影响
- Exploring AI-driven learning assistance in Chinese higher education: A multidisciplinary and regional analysis
- 作者: Yicheng Sun, Hanbo Yang, Hi Kuen Yu, Richard Suen | 发表: 2026-04 (Vol.69)
- 中国高等教育中AI驱动学习辅助的多学科与区域分析
- AI self-efficacy and knowledge graph-integrated generative AI feedback in higher education
- 作者: Yun-Fang Tu, Gen-ping Liu, Gwo-Jen Hwang, Xu-Wei Chen, Xiao-Ge Guo | 发表: 2026-04 (Vol.69)
- 高等教育中AI自我效能感与知识图谱集成生成式AI反馈研究
🐙 GitHub AI/ML 热门项目
| ⭐ Star | 📦 项目 | 🛠️ 描述 |
|---|---|---|
| 84,857 | browser-use | 让网站能被AI Agent操作使用的框架,自动化在线任务 |
| 53,748+3362 | MoneyPrinterTurbo | 利用AI大模型一键生成高清短视频的工具 |
| 51,271+18872 | deer-flow | 字节跳动开源长时程SuperAgent框架,支持研究、编码和多任务处理 |
| 44,362+2617 | minimind | 2小时从0训练64M参数GPT的工具,从头搭建小型GPT模型的完整教程 |
| 43,374+8220 | TradingAgents | 多智能体LLM金融交易框架,整合多个LLM智能体进行金融分析和交易 |
| 25,124 | sglang | 高性能LLM/多模态模型服务框架,高效的注意力机制和推理优化 |
| 22,697 | repomix | 代码库打包工具,将整个代码库打包成LLM友好的格式 |
| 17,000 | NeMo | 可扩展生成AI框架,支持LLM、多模态模型和语音AI的快速训练和部署 |
| 16,160 | Qwen3-Coder | 阿里Qwen代码大模型,支持多种编程语言的代码生成和理解 |
| 15,376+5382 | hermes-agent | 可成长的AI智能体框架,支持持续学习和扩展 |
| 13,207 | TensorRT-LLM | NVIDIA LLM推理优化框架,提供高性能Transformer推理能力 |
| 12,984 | PaddleFormers | 百度PaddlePaddle预训练LLM模型库,集成多种Transformer变体 |
📈 本期趋势观察
- generative AI 相关研究 4 篇
- higher education 相关研究 4 篇
- math 相关研究 4 篇
- LLM 相关研究 4 篇
- academic integrity 相关研究 2 篇
- agentic AI 相关研究 2 篇
- China 相关研究 2 篇
- learner modelling 相关研究 2 篇
🔬 建议深读
Teachers’ doings in the AI era: From transmission to relation
📖 期刊 · Computers & Education | 2026-08AI时代教师角色转型:从知识传递到关系构建,多国学者合作研究…
Trusting AI to detect AI? A systematic evaluation of the reliability and robustness of current AIGC detection tools
📖 期刊 · Computers & Education | 2026-08系统评估当前AIGC检测工具对学生学术作品的可靠性与鲁棒性…
Fair AI in educational predictions: A multi-group fairness approach using reinforcement learning
📖 期刊 · The Internet and Higher Education | 2026-04使用强化学习的多组公平性方法实现教育预测中的公平AI…
The synergy of pedagogical agents and metaphorical design: Reducing psychological distance to enhance video learning
📖 期刊 · British Journal of Educational Technology | 2026-03-26教学代理与隐喻设计的协同效应:降低心理距离以增强视频学习,129名学习者2x2被试设计实验…
From Untamed Black Box to Interpretable Pedagogical Orchestration: The Ensemble of Specialized LLMs Architecture for Adaptive Tutoring
📄 arXiv | 2026-03-25Monolithic LLM拆分为多个专业LLM组合,实现可解释的自适应教学编排…
CyberJustice Tutor: An Agentic AI Framework for Cybersecurity Learning via Think-Plan-Act Reasoning and Pedagogical Scaffolding
📄 arXiv | 2026-03-19基于Think-Plan-Act推理和教学脚手架的网络安全学习Agentic AI框架…
When AI Meets Early Childhood Education: LLMs as Assessment Teachers in Chinese Preschools
📄 arXiv | 2026-03-07针对中国3600万儿童、25万+幼儿园的师幼互动(TCI)评估难题,提出基于LLM的自动化评估系统…
Rewarding How Models Think Pedagogically: Integrating Pedagogical Reasoning and Thinking Rewards for LLMs in Education
📄 arXiv | 2026-01-20整合教学推理与思维奖励信号用于教育LLM训练,提出教学推理奖励的新方法…
Problems With LLMs for Learner Modelling: Why LLMs Alone Fall Short for Responsible Tutoring in K-12
📄 arXiv | 2025-12-28LLM单独用于K-12学习者建模的局限性——为何不足以实现负责任的辅导…
Problems With Large Language Models for Learner Modelling: Why LLMs Alone Fall Short for Responsible Tutoring in K-12 Education
📄 arXiv | 2025-12-28LLM单独用于K-12学习者建模的局限性——为何LLM alone不足以实现负责任的辅导…
📊 简报由 AI 自动生成 | 数据来源: arXiv, 17个教育技术核心期刊RSS, GitHub Trending | 生成时间: 2026-03-29 15:05
💾 数据已存入研究监测数据库: ~/.qclaw/workspace/research_monitor.db