面向个性化学习监测与质量评估的人工智能方法研究与应用

With the in-depth development of education informatization, education will inevitably develop to a new stage in the current network social environment, that is, personalized learning. However, the existing analysis model is difficult to accurately analyze the student's academic level and learning effect from the perspective of personalized learning and lacks effective monitoring and evaluation methods. To this end, using the learning resources, students and teachers as a research object, aiming at improving students’ individualized learning quality, this project conducts artificial intelligence method and application research on personalized learning monitoring and quality assessment to. Specifically, we firstly comprehensively review the general indicators of student development and personalized learning, and establish a personalized learning index system; secondly, we build a networked learning resource library based on knowledge graph; and then study the method that monitors the individualized learning status of students and evaluates individualized learning quality. Finally, the research results are applied to smart education platform in the public service field. The project proposes an artificial intelligence solution for personalized learning and improves the ability of artificial intelligence to assist in personalized learning and extends the range of applications of personalized recommendation systems. It is the extension and concrete practice of information education to intelligent education.

在当前网络信息化教育环境下，教育必然会发展到新阶段，即个性化学习。然而，现有的分析模型很难从个性化学习的角度对学生的学业水平和学习效果进行准确分析，且缺少有效的监测和评估手段。针对以上问题，本项目以学习资源、学生和教师为对象，以提高学生个性化学习质量为目标，开展面向个性化学习监测与质量评估的人工智能方法及应用研究。具体地，首先全面梳理学生发展和个性化学习的通用指标，建立个性化学习指标体系；其次基于知识图谱构建网络化的学习资源库；进而研究监测学生个性化学习状态和评估个性化学习质量的方法。并将研究成果应用于的公共服务领域的智慧教育平台。本项目提出了面向个性化学习监测与评估的人工智能解决方案，提高了人工智能对个性化学习的辅助能力，是信息化教育到智能教育的延伸和具体实践。

本项目针对学习资源、学生和老师等三类教育主体对象，研究面向学生个性化学习与质量检测的人工智能关键技术，进而设计个性化学习系统并加以应用示范。在项目组的努力下，按照研究计划，完成任务书中各项指标：搭建了基于个性化学习评估的支持系统，实现了面向学生个体行为数据的分析和评测，进而通过推荐模型辅助学生开展个性化学习。项目组还搭建了支持个性化学习的学习资源库和试题库，提出了创新的基于注意力机制的新神经网络框架以及用于挖掘多种教育数据源的多种教育学关系抽取框架，构建了教育知识图谱。在理论方法方面，开展了深入的研究，提出针对个性化学习监测的多种人工智能模型和算法，发表了多篇人工智能领域顶级期刊或会议论文，其应用服务数百万学习者，形成较好的社会影响力。