This Science News Wire page contains a press release issued by an organization and is provided to you "as is" with little or no review from Phys.org staff.

NJIT biologist explores ant colonies, human migration with DARPA award

Can ants help predict the complex dynamics of future humanitarian crises, such as when and where large populations might move during disease outbreaks or armed conflicts?

It's a question that has been posed by Simon Garnier, assistant professor of biological sciences at New Jersey Institute of Technology (NJIT), who recently joined the exclusive company of up-and-coming researchers in the nation with a prestigious 2019 Young Faculty Award from the U.S. Defense Advanced Research Projects Agency (DARPA).

The DARPA Young Faculty Award (YFA) recognizes "rising stars" in junior faculty positions at academic institutions across the U.S., providing "high-impact funding to elite researchers early in their careers to develop innovative new research directions" that address national security challenges.

Garnier received his award based on his research proposal titled, "Ant colonies as an animal model to understand the economic, environmental and conflictive drivers of mass migrations"—becoming just the second-ever NJIT faculty member to win the award, and part of a select group of nearly 400 YFA recipients across all scientific disciplines that have been recognized with such an honor since the award was established in 2006.

"I am thrilled to have been selected for such a competitive award," said Garnier. "It is a fantastic opportunity for myself and for my research group to engage in more applied work that can potentially affect real change in how we understand and respond to migration events following natural disasters and armed conflicts."

Tracking the "Swarm"

At NJIT's SwarmLab, Garnier's DARPA-funded project is beginning to explore the intelligent collective behavior of ants, known as "swarm intelligence," to develop computational models that may predict human mass migrations—driven by everything from climate change-related famine and drought, to civil wars, to economic recessions.

"As we've seen with examples like Central America, Syria or New Orleans during Hurricane Katrina, human migrations can be extremely chaotic and involve millions of people moving across geographic or cultural lines," said Garnier. "To anticipate such large-scale migrations, we need a model where we can study the migration behavior of many social individuals at once under disruptive conditions. ... We are establishing ants as that model."

Garnier says there are few better model systems for understanding our migration dynamics than ant colonies, which organize in ways that share distinct characteristics with human transportation networks, traffic organization and collective emergency responses.

"Ants migrate just as we do based on disruptive environmental factors such as food shortages or threats from other colonies. ... When they migrate there is a great deal of information and social feedback primarily transmitted via pheromones and tactile signals within the colony that influences when and where migration will occur," said Garnier. "Like humans, ants share information about the places they've visited with the other members of their group and make use of this social information to decide when and where to move next.

"We can collect all levels of this data in terms of the global state of the population and how it is distributed as well as the individual state of each ant, such as its origin, the origins of the other ants it interacts with, its physiological state and its exposure to threat or other forms of danger, to explore approaches for predicting migration behavior in ways we cannot ethically or practically do with humans."

In the lab, Garnier's team plans to introduce ant communities to a controlled, video-recorded lab environment of interconnected "habitat boxes," each featuring their own designated nesting and foraging areas. Garnier says his team will be able to manipulate conditions of each habitat, such as altering nutrient content of food to simulate food shortages or manipulating habitat temperature to simulate global warming.

Using computer vision algorithms and deep neural networks, the lab will collect data of individual behavior and colony-wide social interactions as the ants migrate box-to-box in response to these stimuli. "For each box we will have cameras constantly recording to capture every interaction. ... We'll feed those images to the neural networks to identify the ants and use our computer vision algorithms to extract behavioral information," explained Garnier.

Garnier will eventually use the data to validate modeling approaches that best identify drivers of a mass migration, detect its onset and predict its outcome.

"Just as we use different models to predict the weekly weather from those to forecast climate change, we want to recommend appropriate models for predicting migrations during an event like Hurricane Katrina versus a conflict-driven migration from Syria," said Garnier. "In the long term, we want to model the complex dynamics of climate change-related migrations."

As a 2019 YFA winner, Garnier will receive up to three years of grant funding and mentorship opportunities with U.S. Department of Defense contacts.

Since his arrival to the Federated Department of Biological Sciences at NJIT's College of Science and Liberal Arts from Princeton University as a postdoctoral researcher in 2012, Garnier has led numerous research projects as head of NJIT's SwarmLab, an interdisciplinary research lab that studies the mechanisms underlying the coordination of large animal groups, such as ant colonies or human crowds, and their applications to complex problems such as the organization of pedestrian traffic or the control of robotic swarms.

Garnier has been the recipient of other awards, such as the doctoral fellowship from the French Ministry of Higher Education, Research and Innovation, and the Fyssen Foundation Postdoctoral Fellowship. Since joining NJIT, his research has been funded by multiple grants from the National Science Foundation, the James S. McDonnell Foundation and the Australian Research Council. He is an associate editor for the journal Swarm Intelligence. Garnier earned his Ph.D. in ethology and M.Sc. in neuroscience, behavior and cognition from the University of Toulouse (France).

Provided by New Jersey Institute of Technology