1 aDeNu Research Group, Artificial Intelligence Dept., Computer Science School, UNED, C/Juan del. Rosal, 16 ... education and open courses. Keywords.
To be published in Proceedings of the 13th International Conference on Artificial Intelligence in Education, AI-ED 2007, Los Angeles, USA.
Supporting Learning Design via dynamic generation of learning routes in ADAPTAPlan a
Olga C. SANTOS a,1, Jesus G. BOTICARIO a aDeNu Research Group, Artificial Intelligence Dept., CSS, UNED
Abstract. The design phase on the life cycle of the eLearning process is currently one of the main bottle necks in adaptive systems. To facilitate this process and reduce the design effort our approach focuses on providing dynamic assistance to some of the author’s tasks that strongly depend on data coming from users’ interactions. In particular, ADAPTAPlan project is developing a system where the learning route of a student is dynamically built from the combination of user modelling and scheduling techniques making a pervasive use of educational standards (IMS). The author is requested to provide simple information about the course structure, pedagogy and restrictions and the system utilize this information together with the user model to generate a personalize IMS-LD course flow suited to that learner. Since the output is given in a standardized format, the course can be run in any standard based learning environment. This approach will be tested on a real course, “How to teach online”, within the UNED’s program for ongoing education and open courses. Keywords. Learning Design, User Modelling, Metadata and Learning, Learning Objects, Learning Activities, Educational standards, Design templates, Adaptive eLearning
1. Introduction Nowadays it is generally accepted that learning should be a personalised process. In this line, some research projects, such as OPAL, OLO and KOD [1] intend to extend existing educational standards to support adaptive course delivery while others, such as EU4ALL, ALPE are addressing students’ individual needs to support eInclusion. All these projects address a critical design issue, which is to determine a priori all the possible situations that may arise during the course execution, including learning materials, pedagogical models, learning styles and learning needs. However, not everything can be specified in advance by the author because unexpected situations appear at run time that cannot be predicted at design time. Furthermore, even knowing everything in advance does not suffice because of the management problems involved, i.e., describing all the existing possibilities and making the adaptation process sustainable over time. In order to cope with these problems, design-time and run-time approaches were combined in aLFanet project in terms of an extensive use of educational standards (IMS-CP, IMS-LD, IMS-QTI, IMS-MD, IMS-LIP). In this approach, the design created in IMS-LD is central in the learning process and the evaluation performed showed that course authors experienced the design phase as a complex task [2].
1 aDeNu Research Group, Artificial Intelligence Dept., Computer Science School, UNED, C/Juan del Rosal, 16, 28040 Madrid, Spain. E-Mail: {ocsantos, jgb}@dia.uned.es –http://adenu.ia.uned.es
A proposal to tackle this problem came from our experience in a previous project on e-tourism. In SAMAP project (TIC2002-04146-C05) planning and user modelling techniques were combined to provide adapted routes to visit a city. Based on this proposal, ADAPTAPlan project (TIN2005-08945-C06-01) was funded to analyse the capability of automatically generate learning routes adapted to the students’ needs by integrating scheduling, machine learning and user modelling techniques.
2. ADAPTAPlan approach To reduce the design effort of the author, we are researching the way to provide dynamic assistance to support this process and ask the author to focus on those elements that require the author’s experience and expertise. This approach differs from other related course generation approaches based on planning [3, 4] or late modelling [5] and asks the author to focus on the learning process in terms of objectives and learning activities, with their corresponding learning objects properly characterized in IMS-MD. They also consider the educational services (i.e., forums, calendars, document storage spaces, etc.) that support the activities and a set of conditions, initial requirements and restrictions in IMS-LD level B. From the student point of view, the system is going to take into account the individual initial knowledge, motivations, learning styles and learning goals, as well as the interaction data gathered and inferred from the students’ behaviour in the course. This information is also specified in IMSLIP. The user model is based on explicit information provided by the student through questionnaires (in IMS-QTI) and/or data learnt from the analysis of the student’s previous interactions with machine learning techniques. With the information known about the student and the description of the learning tasks defined by the author, the system can schedule the tasks and offer the student a learning plan (equivalent to the concept of learning route from instructional design) adapted to the student’s needs. In fact, this plan is specified in IMS-LD, so it can be run in a standard-based learning management system. The individual student and the student group behaviour are to be monitored to feed back the user model and scheduling systems for next runs. Moreover, the author of the course is provided with reports on the students’ performance, which can affect the tasks design, closing the life cycle of adaptation in eLearning as defined in [2]. This approach goes further than others that consider providing the output in a similar structure to IMS-CP [3]. More detailed, ADAPTAPlan approach can be summarized in the following steps: 1. 2.
3.
4.
The author defines initial requirements such as tasks, milestones, restrictions, services and characterizes the course materials. The planner takes these requirements together with the user model of a particular learner and defines a linear plan which consists on a particularized learning route (in terms of IMS-LD) generated for that learner. This learning route is loaded into the course for the learner. Activities, learning objects and services are offered to the learner according to the restrictions specified in the learning design. This IMS-LD is extended to include in the specification all the resources (activities, services, learning objects) available in the course, and not only those included in that particular learning design. This is required in case the
5.
6.
7.
learning route has to be re-planned because the initial planning of the course fails or gets to a stop point (see 5). If the plan reaches an impasse due to: (i) stop point defined by the author (e.g. an evaluation or a temporal restriction), (ii) fails because the learner diverts from the planned route, (iii) blocks in a point of the course, or (iv) the learner cannot perform an activity, the planner modifies the initial plan taking into account the runtime information. From that moment on, the planner (taking into account the full structure loaded in point 4) guides the learner in the course based on the runtime information. When the course ends, the interactions of all the learners are analyzed and used to build a generic IMS-LD. This learning design abstracts all the particularized initial IMS-LD together with the runtime modifications and builds an IMS-LD with all the possible learning routes in terms of properties. Educational Services Questionnaires (IMS-QTI) Material (IMS-MD) Author (design)
Sequence of activities and use of services (learning route particularized for each learner) IMS-LD
Learner 1
Planning engine LMS (dotLRN)
Restrictions (IMS-LD_B)
Runtime feedback
User Model (IMS-LIP)
Interaction data
IMS-LD_B
IMS-LD_A
(generalized for all learners)
(each learner’s nteractions)
Figure 1. General overview of the ADAPTAPlan approach
References [1] Paramythis A., Loidl-Reisinger S., and Kepler J. Adaptive Learning Environments and e-Learning Standards. Electronic Journal of eLearning, EJEL: Vol 2. Issue 1, March, 2004. [2] Boticario, J.G., Santos, O.C.. Issues in developing adaptive learning management systems for higher education institutions. International Workshop on Adaptive Learning and Learning Design, Adaptive Hypermedia 2006. Ireland, 2006. [3] Brusilovsky, P. Vassileva, J. Course sequencing techniques for large-scale webbased education. Int. J. Cont. Engineering Education and Lifelong Learning, Vol. 13, Nos.1/2, 2003. [4] C. Ulrich. Course generation based on HTN planning. Proc. of the 13th Workshop of the SIG Adaptivity and User Modeling in Interactive Systems, pp. 74-79, 2005 [5] Zarraonandia, T. Fernández, C., Dodero, J.M. A late modelling approach for the definition of computersupported learning processes. International Workshop on Adaptive Learning and Learning Design, Adaptive Hypermedia 2006. Ireland, 2006.