可解释学习者模型:可信个性化学习的技术关键

Explainable Learner Models:Technical Keys to Trustworthy Personalized Learning

基于人工智能的个性化学习技术一直是智能教育领域的研究热点,其技术挑战在于如何构建全面精准的学习者模型。在非常注重公平、伦理和责任的教育领域,人工智能的“黑箱”本质可能会阻碍人对机器决策的信任,因此构建透明和可解释的学习者模型尤为重要。通过对学习者模型特征、结构和决策结果的解释,可以让教育关益者理解其动机,接纳其决策,实现更好地人机协作。该研究将可解释人工智能的技术理念延伸至个性化学习中,通过分析其研究现状,阐明实现可解释学习者模型的必要性,并剖析现有的可解释学习者模型实例的技术原理,最后提出可解释学习者模型的基本框架,旨在将可解释性作为学习者建模的关键原则。对可解释学习者模型的研究可为个性化学习系统的设计、开发、应用到评估的整个周期实现可解释性提供借鉴及参考,驱动可信个性化学习成为可能。

Personalized learning technology based on artificial intelligence has always been a research hotspot in the field of intelligent education,and its technical challenge lies in how to build a comprehensive and accurate learner model.In the field of education,where fairness,ethics,and accountability are very much in focus,the“black box”nature of AI may affect the user’s trust in the machine’s decision,so building transparent and explainable learner models is particularly important.By clarifying the characteristics,structure and decision result of learner models,the education stakeholders can understand its motivation and accept its decision,so as to achieve better human-machine cooperation.This paper extends the technical concept of explainable artificial intelligence to personalized learning.By analyzing its research status,this paper clarifies the necessity of realizing explainable learner model,analyzes the technical principles of existing examples of explainable learner model,and finally proposes the basic framework of explainable learner model,aiming to take explainability as the key principle of learner modeling.The research on the explainable learner models can provide reference for the whole cycle of the design,development,application and evaluation of personalized learning system to realize the explainability,and lay a foundation for the realization of trustworthy personalized learning.