Hoping to Spur ‘Learning Engineering,’ Carnegie Mellon Will Open-Source Its Digital-Learning Software - EdSurge News


In an unusual move intended to shake up how college teaching is done around the world, Carnegie Mellon University today announced that it will give away dozens of the digital-learning software tools it has built over more than a decade—and make their underlying code available for anyone to see and modify.

Among the software slated to be released under an open-source license is the university’s pioneering adaptive-learning project, the Open Learning Initiative, as well as a learning analytics platform LearnSphere. Officials estimate that developing the software has cost more than $100 million in foundation grants and university dollars.

The goal of the software giveaway is to jump-start “learning engineering,” the practice of applying findings from learning science to college classrooms.

If it takes off, the effort could result in a free, open-source alternative to a growing number of commercial adaptive-learning and learning analytics tools aimed at colleges. One of the biggest concerns by college leaders about buying such tools from commercial vendors is whether colleges will have access to the underlying algorithmic logic—or whether the systems will be a “black box.”

Carnegie Mellon has a long history of commercializing its digital-learning tools, though. In 1998, for instance, researchers from the university formed a spin-off company called Carnegie Learning that popularized the math tutor software developed at the university. More recently, in 2013, a company called Acrobatiq was formed to try to bring lessons learned from the Open Learning Initiative to the market.

But the standard path of commercializing software has not worked well for the latest generation of digital tools, argues Norman Bier, director of the Open Learning Initiative and the executive director of the Simon Initiative, a learning-engineering project at CMU leading this national effort.

“What we’re seeing is once something gets pushed out into the market, often the things that made it very, very effective aren’t the things that make it sell,” he says. And when software companies make changes that lead to greater sales, he adds, “often you’re compromising the thing that made the innovation effective in the first place.”

He says the hope is that by running Open Learning Initiative software and other teaching tools as open-source projects developed by universities, they can “keep things rooted in the research that’s driving its effectiveness.”

Some of the software that CMU plans to release as open source were already free to download, but until now the source code was not made available. But Kenneth R. Koedinger, a professor of human computer interaction and psychology at Carnegie Mellon, says that moving to an open-source model means that colleges can fully customize it, and potentially share back with other users what works and what doesn’t. “When people think about using free tools coming out of a university, they wonder if they’re using something sustainable,” he says. “One way to build confidence is to say, ‘The source code is going to be available [to review], and we’ll build a community to keep it alive.’”

As part of the new effort, Carnegie Mellon says it will also try a new approach to working with other colleges to encourage adoption of the software. To do that, it has partnered with the Empirical Educator Project, a group of colleges and edtech companies led by e-Literate, a popular education blog, that holds summits and encourages participants to try each others’ tools and approaches. That group says it plans to announce more details about the partnership at its upcoming meeting in early May.

"We need a scientific revolution in education akin to the one that we had in medicine 150 years ago," said Michael Feldstein, coordinator of the Empirical Educator Project, in a statement. "This isn’t a silver bullet, and it isn’t charity. It’s an invitation to the educators of the world for us all to solve big problems together."

Plenty of challenges lie ahead for Carnegie Mellon’s ambitious project.

For one, the effort aims to encourage professors at other colleges to use its software, but it is offering no grant funding or support to make that happen. Having the tools available is important, but changing culture and practice is much harder, says Brandon Muramatsu, associate director for special projects at MIT’s Open Learning project, noting that the training and community are often the most important. “What would be really interesting is if they had donors or foundations make $10 million or some amount available to help universities implement these tools.”

Previous projects similar to what Carnegie Mellon is attempting have struggled to take off. The Sakai Project, an effort to build a community-source learning-management system as an alternative to commercial systems like Blackboard, started with great fanfare and grant support in 2004—but its use has shrunk to less than 3 percent of the LMS market in North America in 2018, according to a report by e-Literate.

Slow-Changing Culture

To understand why Carnegie Mellon is making a stand for learning engineering, it’s important to know who Herbert Simon was.

Simon, who was a professor at CMU and won a Nobel Prize in economics in 1978, was also one of the earliest proponents of learning engineering, and he wrote forcefully about the need for colleges to improve their teaching by applying scientific approaches in the classroom, way before smartphones, Google or even the internet existed.

In a 1967 article in Educational Record, he said that, compared to other organizations in America, colleges are run by amateurs, calling professors “almost completely untrained in the skills of professing: that is, of teaching.”

“We take the traditional organization of colleges so much for granted that we must step back and view them with Martian eyes, innocent of their history, to appreciate fully how outrageous their operation is,” he wrote. “If we visited an organization responsible for designing, building, and maintaining large bridges, we would expect to find employed there a number of trained and experienced professional engineers, thoroughly educated in mechanics and other laws of nature that determine whether a bridge will stand or fall. ... What do we find in a university? Physicists well educated in physics, and trained for research in that discipline; English professors learned in their language and its literature (or at least some tiny corner of it); and so on down the list of the disciplines. But we find no one with a professional knowledge of the laws of learning, or of the techniques of applying them.”

Leaders of CMU’s new effort run out of a center that bears his name and say they are trying to help realize Simon's dream of bringing learning engineers to campuses.

But doing that will mean a change of culture.

“Like Herb Simon said, we need to change higher ed from a solo sport to a collaborative research activity,” said Koedinger.

That’s why leaders are focusing their efforts on building a community of people using the software tools and sharing their practices, says Lauren Herckis, an anthropologist at Carnegie Mellon University who is a faculty member with the university’s Simon Initiative. She has studied the reluctance that many professors have to turning to research to improve their teaching.

Professors are more likely to adopt a new practice if they hear it from a colleague informally than if they read a detailed journal article backed by data. “We need models, we need colleagues, we need collaborators, we need partners,” she says.