It’s no surprise that some major advances in artificial intelligence come from defense research — that’s where the money goes, with over half of America’s research and development spending focused on the military. But as a recent study on treating bipolar disorder from the University of Cincinnati demonstrates, some of that research eventually makes its way back around to more humane causes.
The study used EVE, an AI designed to create other AIs. While this sounds like an apocalyptic scenario waiting to happen to anyone with a taste for science fiction, the researchers responsible for EVE are confident that there’s nothing to worry about, at least as far as Skynet scenarios go. Nicholas Ernest, study co-author and founder of Psibernetix, Inc, says that the algorithm he developed is not a sentient being like the villains in the “Terminator” movie franchise but merely a tool.
EVE uses what the study authors call “genetic fuzzy trees,” algorithms that use fuzzy logic to simplify complicated decisions in a way that resembles learning. Kelly Cohen, a professor in UC’s College of Engineering and Applied Science, compares the process to teaching a child to recognize a chair. Once you’ve seen a few chairs, you know what a chair is, and you can recognize future chairs even if they have very little in common with the chairs you’ve seen.
“We do not require a large statistical database to learn. We figure things out. We do something similar to emulate that with fuzzy logic,” Cohen said.
Ernest originally used this process to develop Alpha, an artificial intelligence that wasn’t designed to recognize chairs. Instead, it recognizes enemy fighters. It’s capable of out-shooting and out-piloting humans in simulations.
“It seemed to be aware of my intentions and reacting instantly to my changes in flight and my missile deployment,” said retired U.S. Air Force Col. Gene Lee after going up against Alpha last year. “It knew how to defeat the shot I was taking. It moved instantly between defensive and offensive actions as needed.”
But military applications only scratch the surface of what Cohen and Ernest believe this technology could be capable of. They brought their research to David Fleck, who was working with UC’s Center for Imaging Research. Together, the three of them put together a team to tackle a thorny medical problem: how to decide whether or not lithium was an appropriate treatment for any given patient with bipolar disorder. The best existing models can only predict who will respond well with a 75 percent accuracy, so researchers put AI to the test to see if it could do better.
It could. It really could. Not only did the LITHium Intelligent Agent (LITHIA) successfully predict how patients would respond to lithium 100 percent of the time in trials, it also predicted the reduction of manic symptoms with 92 percent accuracy. For anyone who’s struggled with finding the right medication for their mental health issues, numbers like that are amazing. LITHIA could well be life-changing for a lot of people if these results continue to hold true. The initial sample was small, with only 20 subjects, so we need a lot more trials before we’ll know for sure.
Dr. Caleb Adler, the UC Department of Psychiatry and Behavioral Neuroscience vice chairman of clinical research, explains the value of this technology from a medical point of view.
“Over the past 15 years there has been an explosion of treatments for mania. We have more options. But we don’t know who is going to respond to what,” Adler said. “If we could predict who would respond better to treatment, you would save time and consequences.” Bipolar disorder is manageable with the right treatments, and LITHIA could help make those treatment far easier to find.
According to its creators, this technology could be applied to other medical problems, too. Cohen believes that the model could help personalize medicine to individual patients like never before, making health care both safer and more affordable. It could lead to fewer side-effects, which means fewer hospital visits, less secondary medication and better treatments. The team is how working on applying it to concussions, and there’s likely much more to come.
If EVE doesn’t create Skynet in the process, it may well make us all healthier and happier in the not-too-distant future.
[Prediction of lithium response in first-episode mania using the LITHium Intelligent Agent in Bipolar Disorders]