The term adaptive e-Learning refers to a set of techniques oriented to offer online students a personal and unique experience, with the final goal of maximizing their performance.
Adaptive e-Learning is based on the principle that every student is unique and has a different background, educational needs, learning style, etc.
The objective of Adaptive e-Learning is being able to capture those diferencies and translate them into contents and training processes which are relevant for each and every individual student.
During the 70s, artificial intelligence burst into education and the Intelligent Tutoring Systems (ITS) were born. Their goal was mimicking human tutoring capabilities, but in an automated and computer based fashion.
A number of ITS systems were developed during the 70s and 80s decades, including SCHOLAR, SOPHIE, GUIDON or WEST.
ITSs systems were mostly experimental and limited when it came to knowledge domains where they were applied.
The world wide web was born during the 90s. New areas of study followed that birth, such as those oriented to find out how to serve web users, relevant and customized contents according to their profile, interests, knowledge ,etc.. This filed of investigation was called Adaptive Hypermedia (AH).
Many training courses that had been developed under proprietary formats and systems were migrated to the web, creating the concept of e-Learning as we understand it today, i.e., accessible through the Internet.
Naturally, two concepts started to fit together. Intelligent Tutoring (ITS) and Adaptive Hypermedia (AH) became the two components of Adaptive Educational Hypermedia (AEH) which is the fundational seed of Adaptive e-Learning.
The architecture of an Adaptive e-Learning System has the following building blocks:
- Knowledge domain: Contains the set of knowledge that needs to be transmitted and taught to students.
- Student model: Contains information about the knowledge, capabilities, preferences, learning style, etc. of each student.
- Tutoring model: Represents the intelligence making decisions about which content elements, execises, materials, etc. are presented to the student in order for him or her to acquire the knowledge contained within the knowledge domain. The tutoring model bridges the gap between the knowledge and the student.
- User interface: Allows for the user to interact with the system and presents whatever contents have been selected by the tutoring model.
Adaptive e-Learning Architecture
Each of this components, as such, is an area of investigation itself:
- How to model knowledge in a way it can be presented in an adaptive fashion?
- Which variables define and represent a user and allow to perform adequate tutoring decisions?
- How are tutoring decisions made?
- Which user interfaces reinforce learning and are adequate for a given user?
The real implementations
There are a number of real Adaptive e-Learning implementations, out there in the market. Each of them implements in a different way the architecture components described above.
There is an important difference among all of them that has mainly to do with their tutoring model, and the technology behind, i.e., how to make decisions about the content to be shown to a give user?
There are two main possibilities:
- Rule based systems, create conditional pathways through the contents of a given course, based on intermediate check points where students show their knowledge level on a set of topics.
- Recommendation based systems, use “machine learning” techniques to infer the different existing user profiles and the learning pathway that will bring each profile into a successful learning experience.
No matter which approach is taken, adaptive e-Learning must be capable of providing a customized online learning experience, making this experience relevant for the student and helping to achieve higher success rates.