Adaptive learning is a dynamic and innovative space. Due to advancements in artificial intelligence (Al) and machine learning (ML), adaptive learning has evolved from pre-programmed learning paths to understanding student learning and focusing on content recommendation through data aggregation. As a result, new platforms include supplemental programs to traditional learning methods—such as chatbots—as well as vertically-integrated, full-stack teaching platforms used at the school and enterprise levels.
The Defense Logistics Agency (DLA) seeks artificial intelligence and machine learning (Al/ML) approaches to adaptive learning in order to support a large number of skilled and unskilled users from agencies across the Department of Defense (DoD). In particular, DLA seeks to understand the current commercial leading-edge technologies that can provide learning on-demand and adjust to learners’ needs in order to improve training outcomes.
Adaptive learning seeks to tailor the learning experience to the skills and capabilities of the learner on an ongoing basis. Adaptive learning systems attempt to do this at scale with the commercial and educational goal of supporting more students without compromising the quality of the education. Until recently, most teaching platforms used decision trees which lead students through a pre-programmed learning path based on right and wrong answers (or if/then statements). These look like they adapt to the student, but it is still just a preset path and therefore highly limited in its adaptability. This method of technology-enabled learning is considered a precursor to the newer, more cutting-edge methods of teaching that use data aggregation and machine learning to adjust what content is delivered, when, and at what speed. The goal is to use this data-centric approach to better take into account how students learn.
Content recommendation is at the core of recent advances in adaptive learning technology rooted in machine learning (ML). Like all ML and many ‘big data’ implementations, adaptive learning ML models operate by ingesting large volumes of data in order to identify patterns and develop strategies. Such algorithms are often embedded in Learning Experience Platforms (LEPs or LXPs). These platforms attempt to quickly surface higher quality and relevant content based on learned associations. These can incorporate a wider variety of media, including podcasts and articles, and often have a “Netflix-like” media library with recommendation algorithms. As these features suggest, this space has benefited immensely from developments in other spaces where content recommendation is relevant, from retail to online media. An LEP can link into multiple different e-learning platforms—or content libraries—and surface the most relevant information for users. For example, a user could log into a single portal to access LinkedIn Learning for a class on Excel, a company’s proprietary content library, podcasts relevant to a current project, and SkillSoft to learn about compliance issues.