The aim, says the books’ creator, is to “explode the book” and rethink how students learn from texts.
“We want to be able to create the perfect book for every person,” says Richard Baraniuk, director of the OpenStax project at Houston’s Rice University, which is behind the books. “Ultimately, we want a system that turns reading the book into an exploration of knowledge.”
OpenStax already offers an array of online and printed textbooks on subjects including economics, biology and history. For the past three years, researchers have tracked how students in 12 US schools use the books in their studies, including information on how they scored on questions.
That work is now being used to train machine-learning algorithms that give OpenStax’s biology and physics textbooks the ability to adapt to individuals. If a reader seems to be struggling with a particular topic – acceleration, say – the book will slot in additional explanations and practice questions, and increase emphasis on related subjects, such as centripetal force, that could otherwise trip that person up.
The adaptive textbooks also incorporate a learning method called retrieval practice, in which material that students have already learned pops up again in occasional quizzes. This method has been shown to enhance students’ ability to retain material, and the algorithmic textbooks will be able to decide when to ask questions based on past exercises.
Digital textbooks are not new – but despite their potential, they have yet to be widely adopted.
“Universities are just not suited for developing and serving such large-scale products. We need start-ups for that,” says Peter Brusilovsky of the University of Pittsburgh in Pennsylvania, one of the designers of the interactive learning system ELM-ART (Episodic Learner Model – The Adaptive Remote Tutor). “If done right, adaptive textbooks could help us to learn faster and better.”
Such personalised learning is designed to give students who are struggling time to understand subjects, while faster learners can surge ahead without getting bored. Software is a great way to do this – at Summit Preparatory Charter High School in Redwood City, California, students spend a portion of each week working independently with a computer program that suggests assignments and tracks progress, but students choose how to spend their time and set their own pace. This approach has helped Summit to become one of the top 20 schools in California, according to US News and World Report.
The initial roll-out of OpenStax in Houston high schools will be relatively small, but large institutions have also expressed interest.
Salt Lake Community College, which has more than 60,000 students and is the largest higher-education institution in Utah, wants to pilot OpenStax’s algorithm-enhanced textbooks next year in political science, business and mathematics classes. Jason Pickavance, director of educational initiatives at the college, says he is curious to see whether the books improve student performance.
“We have such a varied student body in terms of college readiness,” Pickavance says. “What they need is more individualised attention, more tutoring. The courseware has the potential for us to get that mix right.”
Whether the books are successful will depend on teachers, says Ben du Boulay, who works on artificial intelligence at the University of Sussex, UK. They are the ones who will ensure that students make the most of their books – for instance, by working out what to do when the books identify a common problem area among their students.
“If all we needed was books, why have teachers?” de Boulay says. “It’s the educational interactions around private study that make the difference.”
Syndicated content: Aviva Rutkin, New Scientist