自闭症谱系障碍儿童个别教育计划智能推荐与自适应干预

自闭症儿童教育干预是亟待解决的重要社会问题，是信息科学、教育科学、认知科学等学科交叉领域的研究前沿，属于我国特殊教育领域的一项重大需求。本项目面向自闭症儿童个性化适应性干预的基础科学问题，重点突破量化评估、资源推荐和路径规划三个方面的技术瓶颈：针对个体特征及发展轨迹的异质性，提出技术增强的形成性评估方法，实现量化评估和特征模型的准确构建；针对个别教育计划任务重、实施难、成效低，智能化水平不足，构建自闭症领域专家知识库，基于规则和案例推理的混合推荐方法，实现个别教育计划的智能推荐；针对静态化干预方案难以满足个体能力随干预进程动态发展的需求，提出融合深度学习与强化学习的动态路径规划方法，实现个性化干预的自适应。基于以上工作，实质性提升自闭症儿童量化评估的准确性，智能推荐个别教育计划，自适应规划学习路径，实现人工智能技术对自闭症儿童个性化教育干预的赋能，促进特殊教育现代化发展。

Educational intervention for children with autism spectrum disorder (ASD) is an interdisciplinary research frontier for fields including Information, Education, and Cognitive Science, and is a major demand of special education in China. Focusing on the basic science aspect of the personalized adaptive intervention for children with ASD, the current project aims to break through three technical bottlenecks: quantitative evaluation, resource recommendation, and path planning. Targeting children’s individual differences of the characteristics and development trajectory of ASD, the current project proposes a technology-enhanced formative evaluation to achieve characteristics’ quantitative assessment and accurate modeling. In view of the heavy task load, implementation difficulties, low effectiveness, and the lack of intelligence level of individualized education program (IEP), the current program aims to establish the Knowledge Base of ASD that utilizes a rule- and case-based recommendation approach to achieve the intelligent recommendation of IEP. In addition, as the static intervention is insufficient in meeting the demands of individual progresses made during the intervention, the current project presents a dynamic path planning approach that integrates deep learning and reinforcement learning to accomplish the self-adaptation approach of personalized intervention for children with ASD. Taken together, the current project will significantly improve the accuracy of the quantitative evaluation, offer intelligent recommendation of IEP, plan self-adaptive learning path, achieve AI-enhanced personalized educational intervention for children with ASD, and therefore promote the modernization of special education.