AI scholarship at Ohio State runs the gamut in College of Arts and Sciences

As artificial intelligence (AI) grows more sophisticated by the day, educators are finding new ways to incorporate it into their coursework. Its applications go beyond generative language and big data analysis, and faculty at The Ohio State University’s College of Arts and Sciences are demonstrating how AI can improve the human experience. In fields as far apart as visual arts and speech and hearing science, professors are showing students innovative ways to use this technology. Below are just two examples of how AI is influencing research, teaching and learning.

Machine learning quiets background noise

Like many working in speech and hearing science, Eric Healy has been plagued by a single problem: how to reduce background noise for hearing aid users. 

Healy is a professor of speech and hearing science. He also runs the Speech Psychoacoustics Laboratory at Ohio State.

“We’re interested in human hearing,” he said. “How the ear and the brain do their job of decoding and understanding speech. The auditory system does a number of complicated things. Because they happen automatically if a person has typical hearing, we assume they’re simple. But the processes are complex.”

One of the complex processes that Healy has worked to understand is how the ear filters out background noise. Hearing aids have struggled to achieve this goal for decades, he said.

“For 50 or 60 years, there have been algorithms that can remove background noise,” he said. “The problem that came with those is that they’d distort the speech along the way. So, you start with speech you can’t understand because of noise and you end with speech you can’t understand because it’s distorted.”

That is no longer the case. With support from the National Institutes of Health, Healy and his colleague Deliang Wang, a computer science and engineering professor at Ohio State, have created an algorithm that uses machine learning, a type of AI, to suppress background noise in hearing aids.   

“The algorithm is one-size-fits-all,” he said. “It understands what speech is and it understands what noise is. It simply takes away everything but the single, dominant voice in your environment.”

In fact, the algorithm was so successful that some of Healy’s colleagues didn’t believe his results.

“We see five-fold benefit in speech understanding,” he said. “That’s unheard of. It was literally unbelievable.”

Now, though, the consensus is clear. This type of algorithm is the way forward for hearing aids. And the breakthrough couldn’t come soon enough. There is a hearing loss epidemic, Healy said.

“Hearing loss impacts more than people think,” he said. “It impacts professional careers. It can be devastating in terms of social isolation. There are physical and medical consequences to that isolation, including dementia. It’s surprising how prevalent and disruptive hearing loss can be.”

With the use of this algorithm in today’s hearing aids, the devices are more functional than ever. If people are concerned about their hearing, Healy encourages them to see an audiologist.

“The tests are simple, non-invasive and pain-free,” he said. “And solutions are out there.”

Understanding AI in the arts

Chris Coleman is wrapping up his first year as director of the Advanced Computing Center for the Arts and Design (ACCAD) at Ohio State. Prior to becoming a Buckeye, Coleman spent 17 years at the University of Denver.

“I’ve been a professor for a while,” he said. “This is a new opportunity to lead a space that is forward-thinking, has additional resources and has a strong network across campus. It’s an opportunity to operate with a bigger scope – this is a much bigger university – and make bigger things happen.” 

AI is the topic of much discussion in art circles. Concerns abound, from art being copied to artists not being credited for their work, all the way to asking if AI-generated art is art. Coleman is excited to dig into these questions in his new role.

One of his first goals is the creation of a general education course about AI.

“The biggest issue is that AI is sold as this magical black box that just does things,” he said. “I want students to understand what AI does, how it operates and where it comes from. I want to give everyone better tools to decide when to use AI. That’s the essential question: ‘Should I use AI for this project?’”

Students need to understand that behind AI are companies that want to make money off their product, so they will push it as an answer to every question, Coleman said. Instead, students should figure out the places where AI can help them.

“I don’t know how many times I’ve encountered great ideas from students, but the student isn’t comfortable drawing, for example, so maybe they can’t convey their idea to other people,” he said. “I like the idea of a process that can convert something in your head into something you can share with others. AI can be the tool to get from point A to point B but that is not the end of the work.”

Using AI is a skill like anything else students learn, Coleman said. He wants his art students to be able to pair it with something else.

“I encourage students to combine art with other subjects, like AI, to bolster their university studies,” Coleman said. “Developing additional skills as an artist is invaluable, especially in the workplace.”

What AI cannot replace is creative thinking that art students develop over the course of their studies, he said.

“Working in the arts, you learn a nimble, creative problem solving that you don’t find in other disciplines,” he said. “You learn to see other ways around a question. I want students to really value that part of their education.”