A Distance Ecological Model for Individual and Collaborative-learning support

Introduction

Recently, with the development of information and communication technologies, various new teaching subjects and methods using Internet and Multimedia are being introduced throughout the world, and in Japan in particular.  Therefore, now-a-days it is extremely important for a teacher to acquire computer communication literacy (Nishinosono (1998)). So far, there were many system development studies, which aim at fostering and expanding the teachers’ practical abilities and comprehensive teaching skills, by using new technologies, such as computers, Internet, multimedia, etc. In Japan, we developed at the same time communication satellites systems such as the SCS (Space Collaboration System) used as distance education systems between Japanese national universities. In the near future, a teacher’s role will change from text based teaching, to facilitating, advising, consulting, and his/her role will be more that of a designer of the learning environment. Teachers have to constantly acquire/learn new knowledge/methodologies. To help them in this task, we set out to build a free and flexible self-teaching individual environment for them, under the concept of “continuous education”, and at the same time, a collaborative communication environment, to support mutual deep and effective understanding among teachers, by using Internet distributed environments and multimedia technologies. In this environment, a teacher should be able to choose the most convenient learning media (learning form) to learn the contents (subject units) that s/he desires. In this paper, we propose a Distance Educational Model, which is based on the concept of School Based Curriculum Development and Training System, advocated by UNESCO (1998) and OECD/CERI (Center for Educational Research and Innovation), and describe the structure, function, mechanism, educational meaning of this model and finally show some preliminary system evaluation. Moreover, we show a system extension, called RAPSODY-EXT, which embraces the previous system, while adding collaboration facilities.

Distance Educational Model based on RAPSODY

Until now, when a person wanted to take a class on “IT-education”, s/he usually had to leave the office or school.  However, the recent technological developments, e.g., the increased computational power of computers and the increased Internet bandwidth, made it possible and facilitate learning of various kinds of subject contents via a virtual school on the Internet environment. Figure 1 shows how the material on RAPSODY is built by teachers or specialists, via our cell editing and authoring environment, and how the beneficiaries of this material can be various learners, accessing the system remotely from their working place or from their homes.

Figure 1. RAPSODY Usage Image

Figure 2. RAPSODY Integrated Distance Learning Model

The hybrid nature of the learning process in RAPSODY, embracing teacher training models, workplace education models and domain oriented learning models is shown in figure 2.

At the present stage, the RAPSODY system has united the following systems resulting from the research at the Laboratory of Artificial Intelligence and Knowledge Computing:

• a Case Based Reasoning System for Information Technology Education Inoue & al. (1999), that facilitates the usage and sharing of examples cases of actual ITE practices, via a browsing and search module based on case similarity computation;
• a Collaborative Learning Environment based on three Companion Agents, a novice, an expert, and a facilitator Kasai & al. (1999), applied successfully on maths and physics teaching;
• a Hypermedia Navigation system Kayama & al. (2000) based on SOM (Self Organizing Maps) pre-processing and clustering features of the Hyperspace representation, followed by user modeling based on the user history and on a NN (Neural Network) teaching strategy generator, applied on Unix and Hypermedia course contents, and finally,
• a VOD (Video On Demand) -based distance teacher training system, with basic and adaptive feature search Okamoto & al. (2000), allowing teachers to get familiarized with the new developments in ITE.
• In a following step, the system will also integrate:
• a Qualitative Diagnosis Simulator for the SCS (Space Collaboration System) Operation Activity, supporting Mental Model Forming Okamoto & al. (2000b).
• an agent based, adaptive hypermedia distance CALL (Computer Aided Language Learning) system for English teaching Cristea & al. (2000) and
• a discovery learning based CAD (computer aided design) system for learning basics and more about Neural Networks Belkada & al. (2000).

Moreover, the system allows access to a multitude of individual and collaborative learning tools, like a tele-conferencing environment, supporting environments for problem solving, such as Stella, CASE, distance teaching environments, such as Tele-Teaching, and so on.

In the following, we will describe the distance education model we propose.

The Axes

Figure 3. Structure of the Distance Educational Model

Our Distance Educational Model is built on three dimensions (figure 3). The first one is the subject-contents, which represents what the teachers want to learn.  The second one represents the teaching knowledge and skills as well asthe students’ learning activities evaluation methods.  From the third axis, the favorite learning media (form) can be chosen, e. g., VOD, CBR (Case Based Reasoning), etc. By selecting a position on each of the three axes, a certain cell is determined.  A cell stands for a “script”, which describes the instruction guidelines of the learning contents, the self-learning procedure, and so on.  Figure 3 shows the structure of this model. Next, we will explain the meaning of each axis in more details. 

Subject –contents unit

As stated previously, this study focuses on the subject called "Information", which is due to be established as a new obligatory subject in the regular courses of the academic high school system in Japan.  The subject "Information" is divided into three sub-subjects, Information A, B and C.  The specific goals of each sub-subject are listed below.

Information           A: Raising the fundamental skills and abilities to collect, process and transmit “information” using computers, Internet and Multimedia.

Information           B: Understanding the fundamental scientific aspects and the practical usage methods of “information”.

Information           C: Fostering the desirable and sound behavior regarding participation, involvement and contribution in an information society.  This sub-subject focuses on understanding peoples’ roles and the influence and impact of technology, in the new information society.

Teaching knowledge/skills

On this dimension, we have represented three items: sub-subject contents, teaching and evaluating methods for “information” classroom teaching. ‘Teaching methods’ stands for how to use and apply IT, to enhance a student’s problem solving ability.  This involves comprehensive learning activities, such as problem recognition, investigation and analysis, planning and design, implementation and executing, evaluation, report and presentation.  We aim at teachers acquiring the proper students’ achievements evaluating skills, according to each of the above activities.

Learning media (form)

This dimension represents five different learning environments, as follows.

(1) Distance teaching environment (Tele-Teaching): This environment delivers the instructor's lecture image and voice information via Internet, by using the VOD real-time information dispatching function.

(2) Distance individual learning environment (Web-CAI) based on CAI (Computer Assisted Instruction) using WWW facilities: This environment provides CAI courseware with WWW facilities on the Internet.

(3) Information-exploring and retrieving environment using VOD, CBR: This environment delivers, according to the teacher’s demand, the instructor's lecture image and voice information, which was previously stored on the VOD server. For delivery, the function of dispatching information accumulated on the VOD server is used. This environment also provides a CBR system with classroom teaching practices short movies. 

(4) Supporting environment for problem solving by providing various effective learning tools: This environment provides a tool library for performance support, based on CAD, Modeling tools, Spreadsheets, Authoring tools, and others.

(5) Supporting environment for distributed collaborative working & learning based on multi-multi-sites telecommunications: provides a groupware with a shared memory window, with text, voice and image information for the trainees.

“Cell” definition

Table 1. The frame representation of the “cell”

The concept of a “cell” in the Distance Educational Model is quite important, because it generates the training scenario (including information according to the teacher's needs), the subject materials learning-flow and the guidelines for self-learning navigation.  The frame representation of the “cell” is shown in Table 1. These slots are used when the system guides the process of the teacher's self-learning. 

Outline of the teacher training system

The teacher's training environment system configuration is composed of two Distance Education subsystems. One is the training system, where a trainee can select and learn the subject adequate for him/her guided by the script in the “cell”. The other one is an authoring system with “cell” description creation and editing functions. The users of the second environment are, e.g., IT-coordinators or IT-consultants, who can design lecture-plans in this environment.

The training system

The training system aims to support teachers’ self-training (Figure 4). The role of this system is first to identify a “cell” in the model, according to the teachers’ needs, and then, to set up an effective learning environment, by retrieving the proper materials for the teacher, along with the “guide script” defined in the corresponding “cell”. Therefore, the system offers programs for both Retrieving and Interpreting. The training system working steps are:

STEP 1: Record the teacher's needs.

STEP 2: Select a “cell” in the Distance Education Model according to the teacher's needs.

STEP 3: Interpret the “cell” in the guide WM (Working Memory).

STEP 4: Develop the interactive training with the teacher according to the “guide script” in the guide WM.

STEP 5: Store the dialog log-data. The log-data collects information on the learning histories and teachers' needs and behaviors.

STEP 6: Provide the needed and useful applications for the user’s learning activities and set up an effective training environment.

STEP 7: Give guidance-information, according to the “cell” script guidelines, and decide on the proper “cell” for the next learning step. 

 Figure 4. Configuration of Training system

Figure 5. Training system window interaction: first page, message windows, lecture video, message board

Here, it is necessary to explain the dialog mechanism (algorithm) between user and system. The interpreter controls and develops the dialog process between user and machine according to the information defined in our “guide script” description language. This “guide script” description language (GSDL) consists of some tags and a simple grammar for interpreting a document, similar to the HTML (Hypertext Markup Language) on the WWW. The interpreter understands the meanings of the tags, and interprets the contents. An example of GSDL is shown below.

(1)<free>

Definition: description of the text (instruction)

(2)<slot (num.)>

Definition: a link to a slot value in the “cell”

(3)<question>

Definition: questions to a trainee

(4)<choice>

Definition: branching control according to a trainee’s response

(5)<exe>

Call: to relevant “cells”

(6)<app>

Definition: applications used for training activities (e. g., Tele-Teaching, etc.)

Some snapshots of the training system windows are shown in fig. 3. As can be seen, the system conducts an interactive dialogue with the user, to correctly infer his/her needs. The window on the lower left side presents a real time multi-multi site interaction between a teacher giving a lecture and some students.

Authoring system for creating and editing a “cell” description

The system provides an authoring module to create and edit the information in the “cell”.  This module also offers the function of adding new “cells”, in order to allow supervisors (experienced teachers) to design the teachers’ training program.  The configuration of this system is shown in Fig. 6.  The tasks that can be performed by this system are: adding new “cells”, editing the existing “cells”, receiving calls for Tele-Teaching lectures, and managing the lectures schedule.  This system is composed of the “cell” frame creating module, and the “guide script”-creating creating module.  A cell design can be performed as shown in the following.

Figure 6. The procedure of “cell” description for creating/editing of the authoring system

STEP 1: Get the “cell” slot-values: “student’s learning objectives”, “subject-contents/teaching/evaluating method”, “useful tools”.

STEP 2: Substitute the return value of the slot of the prepared media with the training-contents corresponding to the user’s needs.

STEP 3: Substitute the slot-value in the “cell” for the corresponding tag in the “guide script” template.

STEP 4: If “Tele-Teaching” is selected, get information about the lecture, by referring the lecture- and VOD short movie-DB.

STEP 5: Add the new “cell” to the Distance Educational Model.

Figure 7. The authoring system windows: choosing of learning type, inputting fields, creating relational links

The lecture-database consists of “lesson managing files” containing user-profile data, lecture schedules, trainees learning records, lecture abstracts, and so on.  The “guide script” template file contains tag-information, written in the “guide script” description language (GSDL), for all subject-contents items in the Distance Educational Model.

Figure 7 shows some snapshots of the cell editing/creation interface at work. The designer can introduce the tools used, e.g., Netscape, the URL of the new application, the goal, as can be seen in the left window. The upper right window allows learning scenario input. The window below is the relation cell editor, which allows designers to link the existing cells in the database. In the current example, two distinct links have been introduced.

Evaluations and Tests

To inspect the usefulness of our system, we carried out some evaluation experiments. Subjects/ testees were future teaching staff, training in universities, as well as already presently enrolled teaching staff, undergraduate students, and finally, people just interested in information education. From the point of view of practical usage of the system，our goal was to provide the user with a “school” in the actual spot where the teacher/user is，be it workplace or home, and allow access via the various existing network environments. The system usage form can be classified therefore as:

The evaluation of our system can be conducted from two points of view:

We performed an experiment involving as testees 19 teaching staff members and 14 undergraduate students. We classified the testees according to the network speed of the used environment into

The testees had to use the system for a minimum of 2 hours to perform self-training. Moreover, we set up the following experimental conditions for the separate training and designing sessions.

Before enacting the experiment, the distance education modem database contained already 37 training programs; (information A:8 programs, information B:22 programs, information C:7 programs). After more than 2 hours system usage, the testees were asked to fill in a questionnaire that contained 34 questions, organized according to points of view of such as “easiness; convenience “, ”functions” and ”computer environment”.  Except for questions related to the ”computer environment”, evaluation was performed on a linear scale of 0 to 5 points.  In the following. the questionnaire result is presented and analyzed. We prepared 6 questions related to “easiness; convenience “. The global average value was high.  However, the evaluation of the question ”While using the system, did you loose your way?” was low (on average, 2. 9). When analyzing separately the self-training system and the cell construction/editing system, we understood that the evaluation of the latter had influence on the global evaluation. As shown in fig. 5, the cell construction/editing system allows many input items and free descriptions. It seems that it was difficult for the testees to understand the meaning of each item, and then to input the text that conformed to the respective item. Therefore, together with re-analyzing the item names (labels), it is necessary to adopt a design, where it is possible to reference both each item’s description and usage example(s). We have also prepared 12 questions about the ”functions” of the self-training system, as well as 12 questions about the “functions” of the cell construction/editing system.

Firstly, we will present the results regarding the self-training system.  The average evaluation results for the questions ”Did you manage to select a training program that contained novelties for yourself?”, ”Were the message number and contents displayed in the message window and the indication methods appropriate?” was 3. 7, so high. However, the evaluation results of the questions ”Was the number of training programs (candidates) for the next step training activity in the training program list sufficient?” were 3. 2, so relatively low. Moreover, the response dispersion of 0. 17 was, compared to the dispersion of other questions (0. 006~0. 1), high. We think that it is possible to improve this by having more candidate inputs for each training program, and implementing improved training program lists.

Next, we will present the results according to the cell construction/editing system.  The evaluation regarding the question ”By using the connection/ relation editor, did you manage to design links well?” had an average of 3. 5, so was highly evaluated. However, the question ”Did you manage to build scenarios well by using the Guide-Script Editor?” was evaluated with a low average of 2.9. This is due to the fact that the context types defined in the Guide-Script editor were not enough. Moreover, many testees have answered that it was difficult to construct with the editor, as the contexts available for composing scenarios were only 6, and there were many constrains. Therefore, it is necessary to re-analyze the context types. We are currently investigating a description method that allows more degrees of freedom and more available contexts. We have prepared also 4 questions regarding the item ”Computer”. The global evaluation resulting from all questions had an average of 3. 5, so was relatively high and promising.

RAPSODY-EXT: Collaborative Learning Environment Extension

The main collaborative learning extensions of RAPSODY-EXT over the previous RAPSODY system include:

• basic equipment of Synchronous /Asynchronous collaborative learning
• Synchronous /Asynchronous collaborative learning materials development facilities
• Synchronous /Asynchronous collaborative learning support function supplement

The extension of RAPSODY-EXT over RAPSODY can be seen in figure 8. The most important are the collaboration learning support tools, that have to do goal oriented work path planning, to select the tool(s) offering the common working environment, to function as a work history registration/ administration tool, and finally, to do manage results. In this way, RAPSODY-EXT becomes a remote and adaptive educational environment and, at the same time, a dynamic communicative system for collaborative learning and WWW synchronous and asynchronous collaborative learning support. Therefore, additionally to the previously enumerated characteristics of the RAPSODY system, the RAPSODY-EXT extended system also features:

• Synchronous or asynchronous collaborative learning group - or individual portfolio construction
• Collaborative activity logging in the collaborative memory
• Portfolio and collaborative memory knowledge management
• Offer of various directory information

We base many of our management function implementations on one of the strongest tools in collaborative environments, agent technology, which we have not gone into details about in the current paper. Beside of performing low-level management functions and communication functions, agents can build user models, infer interpretations, simulate students or teachers in the collaborative environment, therefore implying different levels of intelligent processing Dillenbourg (1999).

Figure 8. Collaborative Learning environment

Group and individual portfolio construction

In short, portfolio construction takes place as follows. Depending on the collaborative learning style (synchronous/ asynchronous), an individual/ group portfolio is created as a collection of log data about important collaborative activities. Concretely, the following mechanisms are offered: communication message management and knowledge management. The communication management function is a software acting at a higher level than the learner computer terminal and the collaborative learning management server. Depending on the learner’s terminal, the learner data for communication is collected from public and shared applications, is grouped according to the communication message type (data development time stamp, learner ID, message attribute, shared application operation data, etc.) and sent to the collaborative learning management server. On the server, the communication message received from the learner computer terminal is handed over to the knowledge management mechanism. This mechanism does a structure analysis of the message received from the communication message management mechanism, and arranges and integrates the new data with the already accumulated data available in the collaboration learning management database.

Knowledge Management of Collaborative Learning Data

The main goals of the knowledge management in RAPSODY-EXT are to link the information stored in the Collaborative memory, such as the worker/learner group history and the portfolio contents to useful knowledge for each learner; to reflect each learning stage, i.e., to be able to exteriorise not expressed acquired knowledge. In knowledge management, we distinguish between the following two main categories:

Text information management, as in, for instance, concept information extraction: extracted concept dictionary, “on the fly” dictionary; data mining process: computational (frequency, mutual frequency), conceptual (topic/viewpoint, etc.); information visualization: task dependent (word processor, task viewer, etc.), task independent (SOM, state diagram, etc.).

Non-textual information management, as in the mining process via information gain machine learning methods: ID3 (C4.5), decision trees; information visualization: NN usage: SOM, Symbolic “map” generation.

Directory information

The directory information in RAPSODY-EXT has the role to offer information that accelerates group problem solving, as for example: Problem solving tools, Problem solving FAQ, Group work history, Mutual Group matters (information interchange, exchange), etc.

Conclusions and Discussion

This paper proposes a Distance Educational Model called RAPSODY that stands for the networked virtual learning environment based on a three dimensional representation, with the axes 1)subject-contents, e.g., “information”, 2) teaching knowledge and evaluation methods and 3) learning and teaching media (forms). This represents a new framework for teachers’ education in the coming networked age. We have shown the system rationale and explained the architecture of the 3D-representation model training system. Furthermore, we have described a “guide script” language for cell building. Some of the cell contents design represents the research of other Doctoral students and researchers in our laboratory, and was presented at different other international conferences. For instance, a VOD system for classroom teaching video retrieval is being built, for providing teachers with several examples of teaching practice, classified from different points of view and focusing on different aspects of the teaching process. Another example is a CBR system for teachers. Another line of research involves CAI systems, as, e.g., an individualized system for self-training and upgrading in the Neural Networks domain, and an agent-based adaptive system for academic English teaching. The main aim of our system is to support teachers’ self-learning, provided as in-service training. At the same time, we need to build rich databases by accumulating various kinds of teaching expertise. In this way, the “knowledge-sharing” and  “knowledge-reusing” concepts will be implemented. With this system, we can construct various kinds of learning forms and design interactive and collaborative activities among learners. Such an interactive learning environment can provide externalized knowledge-acquisition and knowledge-sharing, via the communication process, and support learning methods such as “Learning by asking”, ”Learning by showing”,“Learning by Observing”, “Learning by Exploring” and “Learning by Teaching/Explaining'”. Among the learning effects expected from this system, we also aim at meta-cognition and distributed cognition, such as reflective thinking and self-monitoring. Therefore, through this system we expect to build a new learning ecology. Finally, we made preliminary functionality tests with the actually implemented system, which have shown us some of the current strengths and weaknesses of the system, and pointed to the future developing directions. For future research, we have to evaluate the system’s effectiveness and usability as a whole from educational point of view.

Moreover, we have presented in this paper the extended RAPSODY-EXT project. RAPSODY-EXT is an individual/ collaborative learning support environment extension of our previous RAPSODY system, and stands for the networked virtual learning environment based on a three dimensional representation. The aim of this system is to support teachers’ self-learning, provided as in-service training, and company employees’ or other learners’ studies, under the umbrella of life-long learning. We have realized the foundation of the integrated distance education project and proposed a Self-Development oriented distance-learning model. This system is superior to a simple rule-based instructional plan, as it allows a better and more natural overview of the global structure, as well as a quick identification of missing parts. Our system is therefore a good example of how to integrate various media and intelligent adaptation techniques with distance learning, in a hybrid, goal oriented manner.

Note:

This paper is based on a previous paper published in CD format and presented at the ISSEI 2000 conference Okamoto (2000).

References

• Belkada, S., Okamoto, T. & Cristea, A. (2000). Supporting Discovery Learning in Building Neural Network Models, 5th International Conference on Intelligent Tutoring Systems, Montreal, Canada.
• Collis, B. (1999). Design, Development and Implementation of a WWW-Based Course-Support System. Proceedings of ICCE99, Amsterdam: IOS Press, 11-18.
• Cristea, A. & Okamoto, T. (2000). MyEnglishTeacher: VOD based Distance Academic English Teaching via an adaptive, multi-Agent Environment. KAIS Journal (recommended submission, acceptance pending),
  http://kais.mines.edu/~kais/
• Dillenbourg, P. (1999). Collaborative Learning, Cognitive and Computational Approaches, Advances in Learning and Instruction Series, Pergamon.
• Inoue, H. & Okamoto, T. (1999). Case Base Reasoning System for Information Technology Education. Proceedings of ICCE99, Amsterdam: IOS Press, 1083-1086.
• Kasai, T. & Okamoto, T. (1999). The Support for Collaborative Learning with Companion Agents. Proceedings of ICCE99, Amsterdam: IOS Press, 760-767.
• Kayama, M., Okamoto, T. & Cristea, A. (2000). Exploratory Activity Support based on a Semantic Feature Map. Adaptive Hypermedia Conference (AH 2000), Trento, Italy.
• Nishinosono, H. (1998). A Teacher Education System for Information Education Teachers at High School Level in the Information Society. Proceedings of JET conference, 523-524.
• Seki, K., Takaoka, R., Matsui, T. & Okamoto, T. (1998). A Teaching Model in Distance Education based on Teacher's Needs. IEICE Technical Report, ET98-12, 49-56.
• Okamoto, T. (2000). The Distance Ecological Model to support Self/Collaborative-learning in the Internet Environment. Paper presented at the the First International Workshop "Developing Creativity and Large Mental Outlook in the Computer Age", International Conference ISSEI'2000, 14 - 18 August 2000, University of Bergen, Norway.
• Okamoto, T., Matsui, T., Inoue, H. & Cristea, A. (2000). A Distance-Education Self-Learning Support System based on a VOD Server, IWALT 2000, Palmerston North, New Zealand.
• Okamoto, T., Cristea, A., Matsui, T. & Miwata, T. (2000b). Development and Evaluation of a Mental Model Forming Support ITS - the Qualitative Diagnosis Simulator for the SCS Operation Activity -, OZCHI 2000, Sydney, Australia.
• Okamoto, T., Cristea, A. & Kayama, M., (2000). Towards Intelligent Media-Oriented Distance Learning and Education Environments, ICCE 2000 Invited Talk.
• Salvador, L. G. (1999). Continuing Education Through Distance Training, Proceedings of ICCE99, Amsterdam: IOS Press, 512-515.
• UNESCO Report (1998). World Declaration on Higher Education for the Twenty-first Century Vision and Action, UNESCO Paris.