Authoring of Learning Objects in Context
{Marcus.Specht, Milos.Kravcik}@fit.fraunhofer.de
Abstract. Learning objects and content interchange standards provide new possibilities for e-learning, nevertheless the content often lacks context data to find appropriate use for adaptive learning on demand and new personalized learning experiences. In the RAFT project mobile authoring of learning content in context has shown the relevancy of contextual metadata for flexible access to learning objects. This paper describes approaches for extending current metadata schemas with context metadata that can be captured together with the assets on the fly and gives them a learning context.
1 Introduction
New requirements for personalized adaptive learning include development of semantic-based and context-aware systems to acquire, organise, personalise, share and use the knowledge embedded in web and multimedia content, achievement of semantic interoperability between heterogeneous information resources and services, and pioneering intelligent content, which is self-describing, adaptive to context and user needs, and exhibits a seamless interaction with its surroundings and the user. This research line addresses the boundaries between knowledge and content, combining new content architectures with emerging knowledge technologies to progress towards context-aware, self-describing and adaptive “atomic” content objects that can seamlessly aggregate to create new content and services, for which the traditional boundaries of different media cease to exist.
Mobile technologies and ubiquitous computing raise new requirements regarding extensions on current standards and exchange formats for contextualisation of resources. The current metadata sets should be extended for capturing and handling additional context data. Authoring toolkits for creating contextualized materials should support contextualized collaboration and live interaction among users performing various roles. One of their primary objectives is to generate as much metadata as possible automatically, based on the current context and by means of various sensors. This will enable more precise retrieval of the data when resources are elaborated by users or provided to learners.
In the last years several initiatives researched scenarios for learning and mobile information support in the classroom. According to [1] the classroom and research in the classroom might be one of the key drivers for a next generation of social software. The classroom gives a variety of scenarios and situations where ad hoc collaboration and the contextualization of information play an important role. In a study conducted in the PEP program [2] 84% of teachers strongly agreed that the quality of teaching was improved by handheld devices in the classroom. New possibilities where seen in the live interaction about data and the reflection about easily exchangeable and copied data sets. In the context of the m-learn project user studies analyzed the different scenarios being relevant at the working context for learning [3]. Most of such studies show a high potential and acceptance for supporting new forms of mobile and contextualized learning approaches in the classroom. From our point of view the integration of focused applications with specialized interfaces and their integration in more complex task contexts are crucial for the design of contextualized learning experiences.
2 Adaptive Methods for Personalization
The major aims of personalized adaptive learning are improvements in effectiveness and efficiency of learning together with higher learner satisfaction. To increase the quality of technology enhanced learning it is important to distinguish what should be adapted, to what features should it be adapted, and how should it be adapted. Additionally to the traditional adaptive factors like adaptive content selection, adaptive navigation support and adaptive presentation, we should consider some new ones, like adaptive learning activity selection, adaptive resource recommendation and adaptive service provision. According to the Adaptive Hypermedia Application Model (AHAM) [4] it is common to base the adaptation process on the domain model and the user (learner) model, possibly enhanced by the goal (task) model, but to provide adaptive services in mobile and ubiquitous computing the context model has to be added. To specify the adaptation itself in a reusable way the adaptation model has to be separated from the domain one (as it is often a case) and in educational settings enhanced by a pedagogical model (more generally it might be an activity or scenario model).
Integrating context modeling and user modeling (Fig. 1) will enable new forms of personalized and adaptive learning experiences. The user and context model specify to what parameters the application should adapt. The main challenges regarding context management include:
- Extensions on current standards and exchange formats for contextualisation of resources
- Automatic acquisition of context metadata
- Understanding contexts and situated cognition
- Creation of tools for development of contextualized applications
- Designing context-based activities involving groups of users interacting within a set of collaborative environments
Fig. 1. Integrating Context Data, User Modeling and Adaptation
3 Mobile Authoring in RAFT
In the European project RAFT (Remotely Accessible Field Trips) [5] the consortium creates learning tools for field trips in schools. The system should support a variety of learners with different tasks either in the classroom or in the field. The main objectives of the RAFT project are:
· to demonstrate the educational benefits and technical feasibility of remote field trips
· to establish extensions on current learning material standards and exchange formats for contextualization of learning material; this is combined with the embedding of learning and teaching activities in an authentic real world context
· to establish new forms of contextualized learners’ collaboration with real time video conferencing and audio communication in authentic contexts
Through the RAFT trials, different phases (Fig. 2) for preparing the field trip, experiencing the field trip in the classroom and in the field, and the evaluation after the field trip were identified. In those phases a variety of stakeholders and participants contribute to the field trip and take an active role in it.
Fig. 2. RAFT Workflow Process
From the prototyping and usage of the RAFT applications by end users we see the following main activities as new qualities of contextualized learning approaches:
· Cooperative task work: The distributed work on a task focuses the interaction and communication between the learners, technology get into the background when the curiosity about the given task and its exploration in physical and knowledge space become the main interest. The context in this sense is an enabling mean that allows the learners to immerse in the learning subject at hand.
· Active Construction of knowledge and learning materials: Users are much more motivated when “self made” learning material get integrated in the curriculum and they have the possibility to extend existing pre-given structures for learning.
· Field Trips are a blended learning process: The different phases of field trips are essential for successfully doing a field trip. Teachers need to specify preparation materials, distribute user roles, and define field and classroom tasks. Additionally after the field trip the collected materials need to be reviewed and archived in standardized formats to ensure reuse and quality assurance.
The RAFT applications support the user with different tools depending on his/her current phase in the field trip process in general: preparation, field trip activity, or evaluation. Therefore different interfaces and widgets give the user access to the LCMS backend system and the live communication channels. The interface and interactions design depends heavily on the type of activity and the degree of personal use or public demonstration. During the field trip the selection of information and collaboration tools is based on the position and current user role. Based on the experiences made in the prototyping phase of the project the implementation of different user roles and interfaces is not based on a software solution for intelligent rendering of interface components but is developed with specialized applications for the different roles and role specific devices for fulfilling the tasks. The RAFT applications can be seen as different components in a blended learning process that is distributed in time, location, social context in the different phases of the field trip.
Our basic architecture in RAFT allows us for creating a variety of widgets using different modalities for input and output. All messages go in a common backend in the LCMS via a web services interface and can be used with different rendering and display widgets. This ensures the most flexibility for communicating between different interfaces in the classroom and the field. Furthermore all clients are notified by a notification service when new messages are available and can subscribe to different communication channels.
4 Extending Metadata Standards
A major focus was the development of contextualized learning materials in RAFT. Beside the traditional learning object metadata (LOM, SCORM) attached to materials in the preparation, field trip activity and evaluation phases, also additional metadata was required for contextualized learning objects. For learners that collected materials it is essential to be able to store information about the location where the materials were collected. For learners exploring a field trip site it is crucial to get information that fits to the current time of the year and the position or maybe even the weather conditions on that day.
Already in an early prototype, called Mobile Collector [6], the learner could annotate a photo (Fig. 3). The photo was shown together with all its metadata. The learner could assign the name and the related concepts (keywords) to the photo, or record audio annotation. Because of the difficulties with text input while on the move, the user could assign the concepts by simply checking them in a predefined list. Based on this manual indexing users could easily find all the photos related to a particular concept.
Fig. 3. Photo Annotation
Later on the RAFT consortium has developed a specialized framework for collecting context sensor data (Fig. 4) in real time together with the learning materials and uses the context metadata to make the collected information accessible to other participants of a field trip. As an example, a learner performing the scout role can collect small pictures or audio annotations and tag them with the location information (sensor metadata) from a GPS device. This tagging and the information instantly appear on the task lists of other team members and are highlighted in the user interface. Traditional learning object metadata can be helpful for adaptive methods on sequencing and selecting the appropriate learning objects for a learner, context metadata enabled new approaches for structuring and accessing shared assets and learning objects.
Fig. 4. Automatic Collection of Environmental Context Data
· Flexible Metadata Schema Support: In the LCMS ALE [7, 8] we integrated a framework to support different metadata schemas and a web based tool that allows authors and creators of learning content to choose from different metadata schemas available.
· Sensor Integration and Sensor Server: Based on the context metadata available on a field trip we integrated the possibility to record sensor measurements and combine them with data collections.
· Context Metadata based filtering and presentation of learning objects: For simple mobile exploration tools based on PDAs or mobile devices we implemented content presentation tools that allow the filtering of information based on contextual metadata.
As one example learners could browse a database of pictures in a biology field trip filtered by the location and the time of the year. Using this approach students could explore and learn about simple questions like “Which flowers grow here at a certain time of the year?”; additionally metadata like the precise time when the picture was taken and the weather conditions on that day can give interesting materials for exploring and learning about important factors of flower growth.
5 Conclusion
The RAFT project raised a lot of technical and interaction issues relevant for the field of designing learning experiences for mobile and collaborative learning. Beside the backend technology based on our LCMS and web services that allows for the combination of different client technologies from electronic whiteboards to mobile telephones, the synchronization and notification of heterogeneous clients accessing a persistent and consistent learning object repository became very important. As we found the field trip a very good example not only the synchronization between different user cooperating on a common task, but also the distribution over the different phases of the field trip (preparation, field trip activity, and evaluation) appear to be an important aspect of nomadic activities for learning and exploration. Last but not least first insights have been gained on the extension of current learning material standards based on the semiautomatic collection of contextual metadata and the combination with assets and learning objects. We are currently exploring the usage of collected contextual metadata for innovative learning experiences and cooperative learning support where learning sequences in the sense of IMS Learning Design can be distributed with the support for different roles in the learning process that challenge the learner for alternative perspectives on learning content.
References
1. Kling, A., Social Software. (2003) Tech Central Station.
2. Tatar, D., et al. (2002) Handhelds go to School: Lessons Learned.
3. Atewell, J. and Savoll-Smith, C. (2003). m-Learning and Social Inclusion - Focussing on Learner and Learning. in MLEARN 2003. London: Learning and Skills Development Agency.
4. De Bra, P., Houben, G. J., Wu, H. (1999) AHAM: A Dexter-based Reference Model for Adaptive Hypermedia. Proc. of the ACM Conference on Hypertext and Hypermedia. ACM (1999) 147-156
5. RAFT Project, http://www.raft-project.net
6. Kravcik, M., Kaibel, A., Specht, M., Terrenghi, L. (2004)
Mobile Collector for Field Trips
In: Educational Technology & Society, 7 (2), 25-33 http://ifets.ieee.org/periodical/7_2/5.pdf
7. Kravcik, M., Specht, M., Oppermann, R. (2004) Evaluation of WINDS Authoring Environment. In: De Bra, P. and Nejdl, W. (eds.) Proc. of Adaptive Hypermedia and Adaptive Web-Based Systems (Eindhoven, The Netherlands), Springer, 166-175 http://www.springerlink.com/index/5151H7EVRREABEAJ
8. Kravcik, M., Specht M. (2004) Authoring Adaptive Courses – ALE Approach. In: Advanced Technology for Learning, 1 (4), 215-220 http://www.actapress.com/onlinejournals/208vol1,%202004/issue_4/208-0824.pdf