About

Wouldn’t it be nice if your textbook learned about you as you learned from it?

The Learning Machines Laboratory (LML) is creating just such an electronic textbook.

Too many of today’s e-textbooks are just as static and uninviting as their paper counterparts. In response, the LML is creating an engaging e-textbook that integrates text, video, simulations, problems, feedback hints, and tutoring and optimizes the learning experience based on each user’s background, context, and learning goals. Our personalized learning system (PLS) couples the latest advances in cognitive science, machine learning, and open educational resources (OER).

Cognitive science: The PLS embodies robust and highly replicable principles from the science of learning that produce superior retention and transfer relative to more conventional ways of studying.
Machine learning: The PLS is flexible, generalizable, and scalable thanks to powerful machine learning algorithms (think Google, Amazon, and Netflix) that optimize the learning experience by analyzing and modeling both the educational content and data from a large number of learner interactions.
Community content development: The PLS leverages the large and growing universe of freely available and easily remixable OER content. Our sister project, Connexions, is one of the world’s first and largest open-access educational repositories. Over 2 million unique users per month from 190 countries access over 17,000 reusable learning modules combined into over 1000 e-textbooks, courses, and articles.

The key features of the PLS are:

Seamless integration of video, audio, and simulation modules to complement text-based materials;
QuADBase, a question-and-answer database that enables the instructor community to develop and share assessments;
Focus, a peer review management system that interfaces directly with Connexions’ lens system to enable quality control of content and professional development for authors and instructors;
Integrated data collection to continuously track each student’s progress through all aspects of a course and feed back analytics to the instructor;
An automated “tutoring engine” that uses advanced machine learning algorithms to close the learning loop by suggesting to the student and instructor “just-in-time” hints and remedial materials for concepts that the student is finding difficult.

The PLS technical infrastructure is extensible and scalable to a wide range of course topics, levels, and instructional modes. Like Connexions, the system is free to use, built on open-source software, and accessible to persons with disabilities. All of the software interfaces (for both content and assessment) are designed to function not just on a conventional laptop/desktop computer, but also on a smart phones and tablets.

The beta PLS system is being deployed and tested at Rice University in Fall 2011. We anticipate better engaged students, more informed instructors, and better learning outcomes using a system that is as effective as today’s intelligent tutor systems but cheaper to build and easier to apply.

The Learning Machines Laboratory is supported by grants from the NSF Cyberlearning Program and Google.