Adaptive Learning Systems: Introductory Discussion

Dr. Fred Saba

By Dr. Farhad Saba
Founder and Editor, Course Strat

The ideal of the industrial age was standardization of goods and services. Mass uniform production of goods brought down the cost of units produced ushering in the economic prosperity of the developed countries in the 20th century. As the United States and other advanced economies are moving into a post-industrial economy, while standardization remains an important standard, it is also becoming clear that the workforce of the future must respond to unique situations with novel solutions; in short, standardized solutions will not be sufficient.

The current system of mass higher education, as Alvin Toffler posited, has a hidden curriculum that prepares learners for an industrial age. It requires that learners sit in classrooms in 45 minute intervals and receive a set curriculum regardless of their prior knowledge of the subject, learning preferences, educational goals or career aspirations.

In response to this new need for education and training a new research literature is emerging that is providing valuable information on how adaptive learning systems can be developed to differentially respond to learners. Also, over the past 10 years researchers in universities, such as, Carnegie Mellon have developed experimental prototypes of adaptive learning systems that are capable of responding to learners on an individual basis. Based on this research pioneers, such as, Dr. Ben Johnson developed several adaptive intelligent tutoring systems and assessment tools that have been in use for several years (Quantum Simulations). Also, textbook publishers, such as McGraw-Hill have developed commercial systems that can assess learners’ initial knowledge of a subject and respond to them individually  (LearnSmart). Learning systems developed by Quantum Simulations are based on the use of artificial intelligence (AI) while systems, such as, LearnSmart rely on a combination of testing the prior knowledge of the learner and leading the learner to the relevant content that s/he needs to improve their knowledge and skills. Other researchers and commercial systems have adopted the use of recommendation engines for creating individualized learning systems. The learner’s experience here is similar to what a buyer experiences when using Amazon.com to purchase a book. As soon as a book of interest is searched or purchased by a user, the recommendation engine suggests other titles to the purchaser. In addition, the expanding use of data collected by major institutions of higher education in their learning management systems or student information management systems to proactively assess and predict the future performance of individual learners is providing another method for making instruction adaptive to the need of the learner.

Research and development on adaptive learning systems is relatively new. There are several models that are being developed by researchers and commercial entities simultaneously. As in other technologies of education that are in their infancy there is no clear winner for any of these models. It can be that an effective system would use a combination of expert systems artificial intelligence, predictive assessment and testing, recommendation engines and big data to make education more relevant to the needs of each individual learner. In this series of articles we will review the research in this field, explain the current models for adaptive learning technologies and systems, examine some of the experimental and commercial products in use today and will provide you with the information that you need to adopt technologies related to adaptive learning for your organization.