Along with colleagues at Princeton University, Cold Spring Harbor Laboratory Assistant Professor Benjamin Cowley has developed an artificial intelligence model that accurately predicts the courtship behavior of real-life fruit flies. In other words, this AI can essentially glimpse into the fly’s future. Putting aside the argument of free will versus predeterminism, this breakthrough may have fascinating real-world applications.
Read the related story: New AI accurately predicts fly behavior
Transcript
Luis Sandoval: You’re now At the Lab with Cold Spring Harbor Laboratory. My name is Luis Sandoval. And this week At the Lab, “Fruit flies’ dating lives.”
LS: Dating is complicated, even for fruit flies. The male fly tracks and woos potential partners by making “music” or vibrations with his wings. But how does he know the right song to sing? How does any animal? How do we?
LS: That’s one thing CSHL Assistant Professor Benjamin Cowley is trying to figure out. More precisely, he wants to know how visual stimuli drive the brain to execute rich and complicated behaviors. And fruit fly courtship is a great place to start.
Benjamin Cowley: Typically in nature, it happens really fast, like 5 seconds. In laboratory conditions, we can optimize it so that it’s not so quick and can last maybe 30 minutes, which is an insanely long time in the lifespan of a fly. This is like a 24-hour date at Starbucks. That gives us a lot of rich data to be able to see this courtship behavior unfold.
LS: Using this data, Cowley’s team built a special AI model of the fruit fly brain that can accurately predict how a real fruit fly will act in response to any sight of a female. In doing so, they found that the fruit fly brain depends on a collection of neurons to process visual information.
LS: Beyond fruit fly dating apps, Cowley’s AI could have curious real-world applications. For one, it could help prevent invasive insects from wrecking farmers’ fruit crops—a big problem out West.
BC: The trick is to release a bunch of sterile males to court with these females and then, when they produce their thousands of eggs, these eggs will not hatch. And that culls the population. So you can imagine if you can make more Don Juans, more of the ladies’ men, you’re going to be more effective at this.
LS: Cowley is hopeful his team’s new AI model will someday help us decode the computations underlying the human visual system. With any luck, it might also reveal the secret to the perfect date.
LS: Thanks for listening to At the Lab. Please subscribe wherever you get your podcasts. And visit us online at CSHL.edu for more fascinating stories like this one. For Cold Spring Harbor Laboratory, I’m Luis Sandoval. And I’ll see you next time At the Lab.