Adaptive e-learning, taken to new levels by incorporating advances in other areas, such as interactive role-playing games, customizable avatars, guided presentation of content, just-in-time knowledge acquisition, CAT (computer adaptive testing), could bring about a total paradigm shift in online education. Imagine a truly individualized educational experience, with an AI mentor (in the guise of an avatar) that cues, requires feedback, research, and interactive just-in-time learning as students replicate the behaviors of successful games. Further, it could be the way to support the 500-student lecture experience, now a staple of bricks-and-mortar universities, without having to go to the labor- and cost-intensive online seminars, usually limited to 25 students, and crippled by low completion rates.
Where is adaptive e-learning today?
Adaptive e-learning is entering a new phase. After being fairly confined to lower-level tasks – randomly selecting content and/or test questions, programming repetitions in “skill and drill” types of activities, especially with voice recognition in language acquisition programs – the new adaptive e-learning programs are being deployed in a number of new venues, many of which involve large numbers of participants in distributed e-learning settings.
The military is using adaptive “smart” systems to help learners practice and adjust as they acquire skills, decision-making abilities, and cognitive self-awareness. Specific examples include the military’s use of adaptive e-learning over distributed systems to prepare soldiers for urban combat.
An example is detailed in J. R. Wilson’s “Adaptive Training,” Military Training Technology: Online Edition. Vol 8: Issue 1, Jan 1, 2003. The article describes how the adaptivity functions in the environment: “Because of the program’s AI, learners can move realistically through their own decision logic chain, where if the blue force element does one thing, the red force responds in an appropriate form. The gamers incorporated an “if-then-else” structure for the AI that makes the best use of the platform’s processing power to provide what Riley terms ‘a tremendous simulation to train leaders to think and react to situations and capture the asymmetric environment we see in our future.’”
CAT (computer adaptive testing) at a new level:
Adaptive e-learning is possible due to recent advances in computer adaptive testing. A leader in CAT is Thomson’s Prometrics, which has presented the following information on its website.
Prometric’s advanced testing technology can present exams as simulated or true-to-life scenarios. Engaging the test taker each step of the way results in a truer testing experience — and a more precise measure of knowledge, skills and abilities. Computer Adaptive Testing (CAT) aims to determine your level of competency by adapting each question to your previous response.
When you answer a question correctly, a more difficult question will appear. If you answer incorrectly, an easier question comes up. Each question’s level of difficulty is based on the answers to previous questions until a consistent level of competency is determined. Adaptive testing technology can establish this competency level more quickly than conventional examination techniques.
Part of the adaptivity of the e-learning environment has to do with using artificial intelligence in building programs that model and even predict a user’s learning styles, based on patterns forged by responses to questions.
Training Place (trainingplace.com) uses an initial instrument (a questionnaire) to “train” the program to begin to attune the delivery and content to the individual’s learning style.
As their website describes it, “The Learning Orientation Questionnaire (LOQ) and other diagnostic instruments are available to assess learning orientation, predict performance, match solutions, and support learning and performance improvements. The Learning Orientation Questionnaire provides scores that are unique indicators of the individual’s approach to learning and describes attributes that characterize learning ability.”
Adaptive e-learning as envisioned by programmers:
Programmers, particularly in aspects applied artificial intelligence, see how the new algorithms can be applied to distributed learning environments. Alex Pongpech writes in
“Adaptive e-Learning Considerations” that “the systems approach adopted here aims to compose e-courses-in a dynamic and adaptive manner-on the basis of learner profiling, using existing database and workflow technologies.”
Other programmers focus on the reusability of the elements. A good discussion of this approach and philosophy can be found here: “A Framework for Adaptive E-Learning Based on Distributed Re-usable Learning Activities,” by Peter Brusilovsky and Hemanta Nijhavan.
Although programmers tend to be enthusiastic, what tends to be missing is a team approach, which some of the innovative providers are assembling. Perhaps the most “leading edge” teams are being assembled by training providers to the military, closely followed by corporate training providers.
Higher education lags behind as community colleges, colleges, and universities tend to rely more heavily on content modules and reusable content objects (SCO’s) packaged and provided by their faculty, instructional design teams, textbook publishers, and independent content providers. The learning model tends to be driven by the learning management system (LMS) – WebCT, e-College, Blackboard, WebTycho, Desire2Learn, etc.
Perhaps it is a case of the cart leading the horse, the tail wagging the dog. At the very least, it tends to restrict learning experiences and place them in what has become the norm – a 25-student section of students, which is highly student- and faculty-time intensive. The prime determinant of quality in this model is faculty interaction.
The adaptive e-learning environment offers a glimmer of hope – it liberates students and faculty from a fairly rigid educational paradigm. Not only that, it allows the possibility of transcending class-size limits, and proposes a truly individualized educational experience, with an AI mentor (in the guise of an avatar) that cues, requires feedback, research, and interactive just-in-time learning as students replicate the behaviors of successful games.
Incorporating avatars in content delivery: online mentor cues and reviews
In order to envision how the adaptive e-learning environment would use avatar-based online mentors, we need to revisit the basics of customizable avatars.
Perhaps one of the best places to explore the way that avatars are allowing individuals to create a persona that expresses an individual’s sense of self is to visit There.com.
There.com allows subscribers to develop a personalized, customized avatar, and then to participate in various chats and scenarios – different “worlds.” The conversations and activities in the chat rooms are not limited by the constraints of a narrative-driven interactive game such as the Half-Life series, Tony Hawk, etc.
The games, however, provide a glimpse of how learning can take place in an avatar-guided environment. For example, when the learner progresses to certain stages or points, the online mentor asks the learner to do certain performative tasks, which could include
---provide responses to questions
---search the online library or on the internet for information that will allow the learner to progress
---ask self-reflective questions
---go to a discussion board and post questions or responses
The goal of the online mentor is to break up the mode of learning and move away from simply reading on the screen to a more interactive one, which taps into other learning styles, such as auditory, kinaesthetic, or visual.
Just-in-Time delivery of knowledge to enhance problem-solving and deeper learning:
At the International Conference of Educational Multimedia held in March 2004 in Quebec City, Quebec, keynote speaker James Gee described how just-in-time learning occurs in interactive, multi-participant, multi-level video games.
Describing the Tony Hawks series of interactive video games which involve creating a persona who progresses through levels of skateboarding challenges, Dr. Gee explained that the learning takes place in a seamless, unforced, and deep manner.
For example, in order to solve problems to progress to the next level, or to acquire necessary skills, the learner must search, retrieve, and understand the information. This is not done in a way that involves rote memorization, or “skill and drill.” Instead, it represents a very productive manner of just-in-time knowledge acquisition.
The most popular of the interactive games, the Half-Life series, illustrates this as well. By creating and maintaining a persona, a profound identification process is set in motion. The learner identifies with the persona, and begins to internalize the worlds, details, and values found in the world. Although this is not precisely a simulation, there is a sense that there is experiential learning at work.
In the case of adaptive e-learning, the key to successful just-in-time knowledge acquisition, is that it is focused on learning goals. Further, the relevance and usefulness of the task as well as the knowledge should be made apparent. Finally, the lesson of interactive video games includes, as Dr. Gee points out, not just projection, but entertainment in a sensory-rich environment.
Attuning content delivery and assessment to learner’s interests:
Using some of the same algorithms, one may be able to develop a decision tree to allow the program to “learn” the students interests, based on his/her responses. Of a menu of 5 or 6 possibilities, the content could then presented in such a way that it makes connections to the learner’s interests, background, experience base, and professional orientation. Attuning could focus on various aspects and could include military, professional avocations, education, hobbies, gender, ethnicity, etc.
After the learner responds to key questions and/or directions, then the content, questions, and assessment would be presented to correlate with the learner’s attributes. This personalization could help with identification, as well as make the content relevant. Both factors could result in a greatly enhanced learning experience, with higher completion rates, and outcomes. Learnframe.com is one such provider of services.
By being flexible, and providing an individualized experience, adaptive e-learning offers significant breakthroughs, which could result in not only a better (more effective) educational experience, marked by deeper learning. It also holds out the promise of a more inclusive, culturally-responsive approach, as well as enhanced access.
If students can enroll in learning “on demand,” and are not constrained by class size, many doors are opened. Further, faculty can spend time developing content, assessing, and interacting with students in guided discussion rooms. It is a way to accommodate large numbers of students, and to perhaps begin to eliminate some of the more ineffective or unrealistic delivery methods (giant lecture hall, etc.).
For those who are concerned that there should be more human interaction in order to provide a chance to mentor and to correct possible misapprehensions, the adaptive e-learning strategy could incorporate a guided, scripted scenario where students, via their own avatars, must enact certain roles and/or accomplish proscribed tasks. In addition, learners can be required to go online and post discussion items in learner forums, and they could write a final paper, if it makes sense with respect to course objectives.