Creating an Adaptive Instructional System for Individualized Learning at Scale

Introduction

We recently discussed the weaknesses in the current approach to virtual education and the potential for collaborative adaptive instructional systems (AIS) to improve outcomes for all learners. By combining computer-based instruction practices with artificial intelligence (AI) and machine learning (ML), educators can personalize the classroom experience at scale. Teachers gain the flexibility to provide targeted instruction. And once created, they can apply these systems across competencies instead of focusing on a single subject area.

The benefits of collaborative AIS are wide-ranging, but we have some hurdles to overcome. Here’s what to consider as you explore creating your own AIS solution.

Is an adaptive instructional system right for you?

Before we dive into the processes behind creating your own AIS, let’s look at some of the potential use cases for adaptive learning. In addition to traditional K-12 and higher education classrooms, AIS is applicable in a broad range of learning environments:

• AI-moderated virtual classrooms
• English learning
• Anti-racism training
• Model United Nations
• Structured team problem-solving
• After action reviews

Essentially, any use case that requires more autonomous learning, better consensus-building, or collaborative problem solving can leverage AIS. Collaborative AIS can also consolidate learning modules from separate systems into a single program that streamlines the classroom experience for teachers and students.

Collaborative AIS case study

Here’s an example of using AIS to create an integrated adaptive learning solution. When the U.S. Office of Naval Research challenged developers to build an intelligent tutoring system for its electrical engineering program, it gave the four “winners” an additional challenge: integrating all of their systems into a single solution. The result was called ElectronixTutor.

ElectronixTutor categorized each topic using knowledge component mapping. The knowledge components list provides a common set of coordinates which ElectronixTutor used to communicate between systems. For example, a student can complete all of the topics required for the “Clamper Function” knowledge component in different systems. Each system then “knows” that the student fulfilled the obligations for that learning component and can offer the appropriate subsequent training.

On the front end, the student and instructor see a single learning environment. Whether the instructor leads the lesson, the system guides the student, or the student takes a self-guided approach, they interact with a single interface.

If someone answers incorrectly, the system can recommend taking another lesson or doing an exercise to improve their understanding of the current concept. Since all user behavior is centralized in a learning record store, the system can learn over time which recommendations lead to the best outcomes and offer up those solutions more often.

Leverage existing tools to build your AIS solution

Some use cases for adaptive learning are relatively specific, and it’s understandable to question if they warrant using all the resources and energy it takes to build a custom AIS solution from the ground up. Thankfully, there are existing services that lower the barrier to entry to adopting adaptive functionalities.

AWS Comprehend uses natural language processing to extract insights from texts or documents, such as key phrases, personal information, tone, or mood. Amazon continuously trains to comprehend using its massive document libraries, so there’s no need to prepare your own training data—significant time savings compared to most machine learning projects.

AWS Lex is a managed service for creating chatbots and intelligent agents. It uses automatic speech recognition and natural language understanding to allow you to build conversational bots quickly. Lex is more mature than many other similar tools on the market because it leverages the same technology that powers Amazon Alexa, which has the advantage of being used by millions of people in a wide range of environments.

Bots are programmed to perform an action based on audio or written responses and can be integrated with your own applications, as well as services like Slack and Facebook Messenger.

AWS SageMaker is a general machine learning service that provides many standard ML algorithms right out of the box, including classification, prediction, and segmentation. It allows you to incorporate ML into your systems more quickly than building a solution on your own.

These mature services offer numerous benefits that accelerate the process of building an AIS solution. They’re secure and encrypted, with automated processes for reporting, monitoring, and logging workflows. They automatically scale, provisioning computing capacity as needed and cutting out a lot of operational overhead. Also, Amazon offers software development kits in multiple languages, allowing you to easily integrate these and other AWS services with the systems you already have.

Overcoming adoption hurdles

Even with the potential of AIS systems to transform educational outcomes, there are significant hurdles to address:

• Developing industry-wide standards—particularly for interoperability—and addressing data privacy concerns
• Convincing schools and teachers that it will make their lives easier and that creating a system is worth the costs
• Retraining teachers and ensuring they receive actionable information from all the data being gathered

Even though these issues are in flux, change is coming sooner than you may think. The pandemic initiated a major shift in virtual learning. Institutions that start to address these hurdles now can get ahead of the curve and work out complications that will inevitably arise in the adoption process.

To better understand the potential of implementing an AIS solution for your institution, explore your options with an experienced education technology partner. To bring these leading-edge solutions to life, you need a partner that understands how to combine existing computer-based instruction practices with machine learning and artificial intelligence to create a customized solution. JBS has extensive experience creating tailored solutions for educational institutions of all sizes and areas of focus. To get started on your custom education software project, contact us today.