转基因学习:构建基于规则、适合大规模招生的e-learning推荐模型

Transgenic learning: towards a rule-based e Learning recommendation model for massive enrollment

目前,教育技术领域的各种模式和教学法没有同时考虑正式学习和非正式学习。推荐系统通常只把正式环境下的活动(比如作业、成绩评定等)作为其输入参数进行储存、跟踪和检索,没有把每一个用户的非正式活动(比如在社交网络上的活动和持续性评估活动)有效结合起来。此外,学习领域的教学辅导系统通常只是建立在内容过滤和其他学生的协作的基础上,这进一步削弱了辅导教师的关键作用。最后,大规模公开在线课程(MOOC)和小规模非公开在线课程(SPOC)已经成为结合正式和非正式环境的教育模式的重要组成部分,在每一个用户的学习路径中发挥关键作用。教育需要通过我们称之为"转基因学习"的破坏性方法促进教与学过程的提升。我们可以通过利用用户的行为和交互信息以及辅导教师高效监控和个性化咨询服务改进每一个用户的学习表现。本文提出一种适合公开和非公开社交网络以及学习管理系统的个性化e-learning推荐模型LIME,这种模型能支持"转基因学习",尤其适合大规模招生的课程和大数据集。文章还详细介绍了根据这个模型研发的框架和软件原型iLIME,以展示LIME模型是如何独立于学校所使用的学习管理系统运行的。文章最后介绍了一个案例,该案例是以面向全校实施一门慕课教学为背景在Apereo Sakai CLE 2.10-svn环境下运行这个模型。文章还讨论了技术问题和挑战,并提出解决方案,目的是为了能实现在真实学习环境下运行iLIME,向学习者提供基于LIME的推荐。

Current models and methodologies in the field of educational technology do not involve engagement between formal and informal learning.Usually,only activities within a formal environment(i.e.,assignments,grades,etc.)are stored,tracked and retrieved as an input parameter in recommendation systems.There is normally no useful combination with the informal activity of every user(e.g.,social networks and continuous evaluation).In addition,tutoring systems in academic domains are usually based only on content filtering and collaboration from other students,which contributes to dissolving the crucial role of the tutor.Last,MOOCs and SPOCs have become a crucial part of educational models combining formal and informal settings,playing a key role in the learning path of every user.Education requires a disruptive approach to boost the learning-teaching process.We call it transgenic learning and by making use of information about the user’s behavior and interactions as well as efficient monitoring and personalized counselling by a tutor,we can improve the learning performance of every user.This paper presents LIME,a personalized e Learning recommendation model for public and private social networks and learning management systems,which supports this approach,specifically for massive courses and large data sets.It then elaborates on a framework and software prototype(i LIME)which has been developed to demonstrate how the LIME model could operate independently of the learning management infrastructure in use.Finally,it reports on a case study developed around the Apereo Sakai CLE 2.10-svn,in the context of a MOOC strategy to be implemented at university level.Technical issues and challenges are also discussed and solutions are proposed in order to run i LIME and deliver LIME-based recommendations to learners in a real academic scenario.