The Complexity Engine
Data analysts are always conjuring the next algorithm to materialize marketing gold. What about an algorithm to help teach kids to read? That’s what three partners asked at the University of Alabama in Huntsville. And they have a good start on the answer.
UAH education professor Philip Kovacs, left, and Ryan Weber, director of business and technical writing at UAH, created an academic search engine designed to increase students’ thirst for knowledge.
Politicians, teachers, social workers, administrators and parents are all looking for new ways not only to improve student performance but also reduce costs.
A new Internet search engine being developed by University of Alabama in Huntsville professors Philip Kovacs and Ryan Weber, in collaboration with Georgia Institute of Technology student Tripp Roberts, is beginning to pique the interest of school administrators and educators across the South and beyond for its ability to do both.
The Complexity Engine is designed to replicate textbook learning by sifting through Internet sources using a complex algorithm developed by Roberts. However, the Complexity Engine transcends what can be gained from a traditional textbook by making careful adjustments to cater to the user’s reading level.
“Let’s say that you’re a struggling reader, maybe you’re a sixth grader reading at a fourth grade level,” explains Kovacs, who teaches in the UAH College of Education. “What the Engine does is slowly increase the complexity level of the material given.”
In essence, a teacher can set certain learning parameters for each student and change those parameters as the student begins to master the material. This sliding scale of difficulty allows the Engine to engage with the student, gradually nudging him or her to greater levels of reading comprehension.
This fresh approach will not only help struggling readers but also more advanced readers who find textbook material boring or unchallenging, the developers say.
“The writing in a traditional textbook tends to be towards the middle,” says Kovacs. “In a traditional classroom, you have kids who are below the middle and kids who are above the middle. This will be more specific to each of them. This will work for anybody wanting to learn anything.”
It’s an innovative approach that addresses a critical problem, particularly in Alabama. In a world replete with Tweets, Facebook statuses and text messages, it’s easy to overlook the fact that illiteracy still poses significant challenges for 21st century educators. In 2003, the Institute of Education Sciences estimated that 15 percent of Alabamians were illiterate, while the Literary Council of West Alabama places functional illiteracy as high as 25 percent.
Kovacs is alert to this need, and the idea for the Engine stems from a project he did for UAH in which he was asked to spend 14 weeks with at-risk learners to try and find new ways to further their reading comprehension. Kovacs wondered if perhaps allowing the students to read what they wanted, instead of something forced upon them by their teacher, would be helpful. He gave students a wide breadth of reading choices and had them give presentations and write blog posts about what they were learning. By the end of the 14 weeks, the control group’s skills had improved by roughly 10 percent, while his experimental group’s scores had increased a remarkable 59 percent.
Still, Kovacs saw room for improvement.
“One of the challenges was that sometimes students would get something that’s just way too difficult; they spent a lot of time wading through context that just wasn’t great,” he explains. “The ‘aha’ moment was when we said ‘Wouldn’t it be cool if there was an algorithm that could do some of that for them.’”
That’s when Kovacs began looking for designers to turn his concept into a practical reality. And that’s when he met Roberts.
“I met Mr. Roberts through LinkedIn, and also through an organization called Appleton Learning that I do consultation for. I asked him if the Engine was possible, and he said it was. He’s been indispensable since then,” Kovacs says.
Kovacs also got in contact with Weber, a professor in the English Department at UAH. “I liked his character, he was someone I thought I could easily work with, and he’s done a great job as far as technical writing is concerned.”
“I was extremely interested in the project from the onset, especially when Dr. Kovacs described it as ‘the Pandora for education,’” says Weber. “I believed that the project had the capacity to be a game changer, and in the two years I’ve been working on it, I’m seeing that possibility come to fruition.”
Over 10 months, the three grew the Engine from a vague concept to a fully-functioning prototype. After this prototype was developed, the Engine underwent strenuous testing in several volunteer school districts.
Teachers who tried it told the developers that they liked the “straightforwardness” of the Engine and its extensive database of useful sources. Students described the Engine as “better than Google.”
“The results were mixed,” Kovacs cautions, but he also notes that the newer models of the Engine have expanded on the database and brought in more varied and pertinent sources.
Kovacs and his team recently have received a grant of $35,000 from Alabama Launchpad, an organization that supports promising start-ups, but Kovacs is still hopeful that more investors will be willing to lend their support to the project — adding that investors would need to be willing to be “change-makers” in the field of education.
Kovacs is keen on implementing a “host of new features” that should augment the experience offered by the Engine and predicting that the proposed features could “do away with standardized testing as a mechanism of educational growth.”
Drew Rutens interned at Business Alabama. He is a student at Hendrix College.