数字技术迭代加速，引发世界范围内的“数字变革”与“学习型社会建设”，揭示了数字时代更加注重因材施教、实现规模化教育与个性化培养的按需结合。文献分析表明，按需学习是实现规模化教育与个性化培养的有效手段，是应对我国未来教育教学均衡发展的重要策略。作者在博士后期间围绕数字时代的按需学习范式这一主题完成了一系列科研项目工作和学术研究，本文是对这些工作的总结。首先，本研究基于知识观、学习观、课程观及教学观的嬗变，结合文献重点阐述按需学习范式的适切性及要素表征，从数字资源、干系人及智能技术视角描述数字技术赋能的按需学习场景，试图回答数字时代背景下，为何学习、学习什么以及如何学习，强调学习过程的“按需适配”，助力学习者完成个性化学习路径（研究1）。而后通过两项实证研究，以期提升数字时代互联网环境中学生学习体验的可调控性（研究2），以及对提升数字时代对学生的学业参与度、学习倦怠的可解释性（研究3）。研究2采用滞后序列分析（LSA）法对学生学习行为日志数据进行了跟踪分析，探索了学习者的人格差异在学习过程中如何影响学生的自主性学习行为，发现通过结合其人格特征采取适应性的导航模块举措、提供个性化的支持能够提升其学习体验与成果，据此可以通过有关基于学生性格的导航行为差异，为自动人格建模领域提供思路。研究3探索了学生的学习参与度如何因社会情感技能（即社交参与、好奇心、孤独感、归属感、韧性和坚毅）和手机依赖之间的相互作用而变化，发现更高水平社交参与、韧性和坚毅的学生比倦怠更容易参与；相反，感到孤独和好奇的学生更容易感到学业倦怠。最后，以Diana Laurillard教授提出的教学范式会话框架与设计原则为基础，构建以知识图谱技术为例的四个典型教学活动序列设计，从教师—学生、学生—学生、学生—学习环境三个视角开展按需学习范式的互动层次研究，实现针对学习需求的干预策略匹配，形成精准学习干预的完整链路，构成个性化学习服务的闭环（研究4）。同时，研究总结了按需学习范式的实践需要面对的挑战和影响学生学习体验、学业成果的关键因素，这些发现为后续相关的研究提供了可能方向。

面向数字时代的按需学习范式设计是推动未来教育发展的重要策略，也是实现教育公平发展的有效手段。在博士后有限的时间里，本研究仅从影响学习者学习行为的因素角度对按需学习范式这个宏大蓝图进行了细微切入点探索，“如何按需、何时按需”这一概念是相对复杂多变的，因此本研究在深度和广度上存在局限性，未来有待进一步拓展探索这一议题，尤其站在学习型社会视角如何通过数字手段去持续提升学习者学习投入度、内驱力和学习成效方面，助力其实现自我价值，研究之路任重而道远。

The iterative acceleration of digital technology has triggered a "digital revolution" and "learning society construction" worldwide, revealing the need for a combination of mass education and personalized training that emphasizes individualized teaching in the digital age. Literature analysis shows that demand-driven learning is an effective means to achieve mass education and personalized training, and is an important strategy to address the balanced development of education and teaching in China's future. During the postdoctoral period, the author completed a series of scientific research projects and academic research on the theme of the demand-driven learning paradigm in the digital age. This article is a summary of these works. Firstly, based on the changes in knowledge, learning, curriculum, and teaching perspectives, this study combines literature to expound the appropriateness and element representation of the demand-driven learning paradigm, and describes the demand-driven learning scenarios empowered by digital technology from the perspectives of digital resources, stakeholders, and intelligent technology, attempting to answer why to learn, what to learn, and how to learn in the context of the digital age, emphasizing the "on-demand adaptation" of the learning process to help learners complete personalized learning paths (Research 1). Then, through two empirical studies, the controllability of students' learning experience in the digital age Internet environment is improved (Research 2), and the interpretability of students' academic participation and learning fatigue in the digital age is enhanced (Research 3). Research 2 uses Lag Sequential Analysis (LSA) to track and analyze students' learning behavior log data, exploring how learners' personality differences affect students' autonomous learning behavior during the learning process. It is found that combining adaptive navigation module measures and providing personalized support based on their personality traits can improve their learning experience and results. Based on this, it can provide ideas for the field of automatic personality modeling through differences in navigation behavior based on student personality. Research 3 explores how student engagement varies in response to socioemotional skills (i.e., social engagement, curiosity, loneliness, belongingness, buoyancy, and grit) and phone dependency. It is found that students who reported a higher level of social engagement, buoyancy, and grit were more likely to be engaged than burned out. In contrast, students who felt lonely and curious were more likely to be burned out. Finally, based on the teaching paradigm conversation framework and design principles proposed by Professor Diana Laurillard, four typical teaching activity sequence designs are constructed, taking knowledge graph technology as an example. The interactive level research of the demand-driven learning paradigm is carried out from the perspectives of teacher-student, student-student, and student-learning environment, realizing the matching of intervention strategies for learning needs, forming a complete chain of precise learning intervention, and forming a closed loop of personalized learning services (Research 4). At the same time, the research summarizes the challenges faced by the practice of the demand-driven learning paradigm and the key factors that affect students' learning experience and academic achievements, providing possible directions for subsequent related research.

The design of the on-demand learning paradigm for the digital age is an important strategy for promoting the development of future education and an effective means of achieving equitable education development. In the limited time of the postdoctoral period, this study only explored the on-demand learning paradigm from the perspective of factors influencing learners' learning behavior. The concept of "how and when to meet the demand" is relatively complex and variable, so this study has limitations in terms of depth and breadth. In the future, further exploration of this issue is needed, especially from the perspective of a learning society, how to continuously improve learners' learning engagement, intrinsic motivation, and learning effectiveness through digital means, and help them realize their self-worth. The research journey is arduous and long.