Anika Puri, of Chappaqua, New York, was on a trip to India when she learned elephant poaching is still a big problem. Determined to help, she built ElSa (short for Elephant Savior), low-cost machine-learning software that can detect and prevent poaching. Puri, 18, won $15,000 from this year’s Regeneron International Science and Engineering Fair for her work.
Q: What’s wrong with current poaching detection methods?
A: There are drones surveying the area in African national parks, but the footage is manually looked at. That’s tedious and error-prone. They’re trying to automate the process with algorithms, but they’re looking at the shapes of objects, which results in very low accuracy.
Q: How is ElSa different?
A: I studied the movement patterns of elephants and humans. I also noticed that elephants usually travel in herds, while poachers tend to travel in smaller groups. Using that, I was able to increase the model’s accuracy of identifying a poacher to over 90 percent.
Q: How would you like to see ElSa used in the future?
A: I hope to be able to implement this sort of methodology in national parks in Africa and Asia. And another big step for me would be to apply this methodology to the conservation of other endangered animals.
(This interview was edited and condensed for length and clarity.)