|Information Technology home | About Information Technology | Courses | Staff | Students | Research|
The PIAVEE project
A Platform Independent Agent-based Virtual Educational Environment
The structure of electronic educational delivery is determined by the available software tools. Attempts to develop open and flexible virtual educational systems have been constrained by a lack of both a software model and effective software implementations.
Many current educational systems have some degree of reliance upon a particular computer platform. This varies from the total system being Operating System dependent through to certain core software elements creating restrictions on the extent to which materials can be shared. Making an educational environment platform independent has the advantage that changes in propriety systems will have minimal impacts upon the functionality of the system.
One of the areas of software engineering that can provide the flexibility and dynamic structure required by e-learning and virtual educational approaches is that of mobile and pedagogical agents. Agents can be created that have varying levels of artificial intelligence to provide organising and management services across a complex educational environment, whether that environment is real-world, internet oriented or virtual.
The primary objective of the project is create a prototype of a virtual education environment using agent technology as the management system. Related to this is the objective of testing out the practicality of a delivery system that uses educational objects or resources as it core data and then provides the teacher with the tools to structure these resources in pedagogy independent ways.
The project will produce software that will be able to collect and classify educational materials as well as being able to organise this material in a teacher-driven mode.
This software will be designed to be used by any computer-literate teacher in any discipline. The software will also be usable by any computer-literate student.
The Agent Based Approach
An architecture, capable of implementation, that can provide a platform independent educational environment would, then, be best based on a version of agent technology, intelligent agent technology including mobile agents for specific components of the architecture.
The effectiveness of this type of system is very much dependent upon the way in which data is stored and the mechanisms that are able to be implemented to keep the information base up to date. A combination of physical data bases and virtual data bases would be implemented in this instance. In part, the system would be design to allow for user definition of the type of primary data storage to use but there would be elements of the system that would have to be based upon virtual data if the agent technology was to be used most efficiently. The advantage of the virtual data base is that it reduces the need for local storage, it allows for rapid management and it can be readily linked into various educational modules.
This is further enhanced through thinking in terms of developing a virtual data base that does not have to store materials but simply points to data items.
Figure 1 presents a schematic for the overall system. The top level of the diagram show the teacher and student users using the Teaching or Learning Resource Agents. The teacher will develop a curriculum structure through the Teaching Resource Agent, creating a local repository. This local repository can be a mixture of data represented as meta-data structures and data stored only as links or pointers (a virtual data base).
The definition of this Local Repository is used by the Information Distribution Agent to extract from the main Meta-data Repository or to allocate links and pointers based on the Virtual Data Repository. The contents of the central data bases are managed by the Repository Management Agent. It works in real time to maintain the integrity of the contents including checking that live links and pointers are valid - for example eliminating the Error 404 when the browser cannot find the link. It sends information to the Content Search and Management Agent that does the "leg work" of collecting data and establishing the links and pointers.
The Student will use a Learning Resource Agent to tap into the curriculum that has been defined by the Teacher and will be able to access the information that is stored or pointed to within the Local Repository.
An important point in understanding the pedagogical significance of this diagram is that the Data Sources shown at the bottom of the diagram can be anything. A particular data source could be standard library catalogue entries; it could be a reading list; it could be an actual set of notes stored in the teacher's computer; it could be a piece of simulation software that the student can run; or it could be a set of links to information on the Web. Any of this material can be updated or modified by a simply reference through the Teaching Resource Agent. Over and above this is the fact that the Teacher is able to generate any form of curriculum based on most current pedagogical models ? a constructivist approach, a formal conventional approach or a curriculum based purely on a system of educational objects. The Learning Resource Agent can instructed to provide access to educational materials in any way the teacher wishes to define.
Perhaps the greatest advantage of this system is the way in which it is independent of educational delivery issues. The teacher is not constrained by the underlying pedagogy of the developer; the teacher can build any type of curriculum system he/she wishes from a traditional sequential term-based curriculum through to one that is based upon an educational resource structure through which the student can be guided.
Any such system has a potentially long learning curve. A clear advantage of an agent-based system is the artificial intelligence that can be built into to it to provide help to a user. The help system can be refined, in real-time, as the system is coming into operation.
The Ethical Issues
Agent technologies, by definition, operate in the background. They are hidden from the user and the user often does not know that an agent is operating. For example, the virus scanning software involves a type of agent that sits in the background scanning those files that it has been told to scan. Its level of real time reporting depends upon the settings chosen by the computer user.
It is important to recognise that the intelligence of the agent is very limited. Agents can "learn" based upon the sophistication of the software and, consequently, the capacity of the computer system to handle the software running in the background. The invasiveness of a piece of software is dependent upon the attitude of the software developer.
Ethical programming of agents is an important issue and CERG will be looking to explore this as a part of the development process.
Most attempts to develop complex educational environments are faced with the continuing costs associated with maintenance. The very nature of physical systems means that someone has to monitor, manage and upgrade the data being handled. The CERG group attempted to build a low-maintenance data base of educational best practice during the second phase of the AUTC funded ICT project. What was recognised was that, even though it did achieve the most of the intended aims, the functionality of this data base was dependent upon those who deposited materials being willing to continue to maintain their entries.
Where a virtual or meta-level data base is generated under the system being proposed here, a key feature would the use of a dynamic indexing system. That is, indexing of data items would be carried out through content searches and/or key word components. There would be need for a limited skeletal framework within which this dynamic indexing could take place. Maintenance of such a system is based upon the way in which appropriate agents are programmed during system development.
It should also be noted that the agent-based management of the system means that there can be real-time maintenance of broken Web-links. This approach will be able to, in real time, establish all links as live links.
As the system will be platform independent, the need to upgrade the software to take into account operating system and software changes is minimised. The expensive process of maintaining version compatibility will almost disappear.
The conceptualisation of distributed educational systems and the application to virtual environments has been discussed for at least 10 years, most notably within the New York City Eiffel Project. The operationalisation of such thinking has been restricted through the availability of a software system that readily allows for complex educational structures such as virtual libraries of educational objects.
The agent-managed educational system being proposed here has the ability to create flexible learning systems in a way that no existing system can. The design philosophy allows for curriculum development ranging from traditional highly structured systems through to highly flexible systems geared towards the self-regulated and independent learner.
A consequence of this is that it is ideally suitable for the development of transportable curricula across international boundaries. It is already recognised that the development of effective distance/e-learning systems is constrained by the tendency to emulate the classroom. This system encourages teachers to look outside that constraint and can be used to implement learner-driven "curricula".