For Simon Garnier of the New Jersey Institute of Technology, traffic on the New Jersey Turnpike has become an opportunity to ponder human inefficiency. Garnier is studying ants and slime mold in order to better understand collective human behavior.
Can ants and mold slime explain collective human behavior?
For most people, getting stuck in a traffic jam on the New Jersey turnpike is a grueling lesson in futility.
But if you're Simone Garnier of the New Jersey Institute of Technology, it's an opportunity to ponder how we became so inefficient compared to other species.
Garnier is studying ants and slime mold in order to better understand collective human behavior.
When you look over an ant colony, and especially if you look up close, it looks extremely messy.
There doesn't seem to be a particular organization to the system.
However, if you take a step back, you're going to see trails that are all coming from one region that branches out, and they seem to be branching out at particular points.
You're going to see the traffic.
If you look at the traffic from far, it is not actually completely random.
It follows a particular path.
You're going to see an actually very well-organized superorganism.
This is Professor Simone Garnier of the New Jersey Institute of Technology, and his research into ants and slime mold has given him a unique perspective of our own collective behaviors.
Every time I'm in traffic, it's like occasion for me to observe the behavior of the other people and try to understand why we are so bad at organizing our traffic when ants, for instance, are so efficient at it.
The work we do is all about how a large-scale organization emerges from the actions of a lot of different individuals.
For example, take a swarm of South American army ants.
Most ant species are going to either dig into the ground to establish a network of chambers.
But the army ants don't do this.
They keep moving all the time in their colonies.
And they're capable of doing this with very tiny brains.
There's no boss in the colony.
There's nobody telling them, 'Well, it's time to do this,' or 'it's time to do that.'
Decisions are communal and often made through the use of pheromones, molecular dollops secreted by each ant as they move along.
That's their main mechanism of communication, main mechanism of orientation inside the rain forest.
But they are practically blind.
And just by using pheromones and touch, army ants can solve a myriad of dilemmas.
If you look at the ground floor in the rain forest, it's extremely messy.
There's a lot of ups and downs, a lot of gaps.
So, these army ants have this very particular behavior adaptation to create structures, ladders, and bridges.
To expedite their traffic.
After all, why go around a gap when you can build a bridge over it?
So, what we do is we build setups that we can place very quickly on top of a trail to force the ants to walk on top of our setup, and then we can record what they're doing with three different cameras.
To figure out what triggers the creation of these living structures.
What we suspect is when an ant arrives in front of a gap, it has to slow down, naturally.
And because it slows down, the ants behind them are piling up and actually start walking over the ants that are slowed down.
And it seems that the signal when someone is stepping on your back is a signal for the ants to stop moving.
Repeat that signal over and over in front of a traffic jam, and you get the beginnings of a bridge -- one even able to adjust itself to shorten the overall path.
And they're capable of doing this with very tiny brains.
But maybe this doesn't impress you.
We big-brained creatures work together to make bridges all the time.
We use apps to tell us the shortest route.
So, then, how about a creature with no brain at all?
So, slime mold is this unicellular organism that's actually a collection of nuclei inside.
So, it's like the cells of your body had all fused together to form this sort of gelatinous blob.
A blob capable of solving one of the most complex problems in sociology and economics, known as the multiarmed bandit problem.
It's a reference to the arm bandit in the casino.
And you might have 50 of these slot machines in front of you, and they don't always have the same reward rate.
But you won't know which machines are better unless you spend hours pumping them full of quarters.
And so, there's gonna be a point where we need to make a decision and stop, because otherwise we will just drain all our resources.
And so that's the multiarmed bandit problem.
You don't have to go to the casino to experience this problem.
You experience it all the time.
If you want to buy a new laptop on Amazon, you don't have all the time in the world to focus on getting all the information, opening all the information, all the options.
It's the same thing in traffic.
You have multiple options when you want to go from your work to your home.
Now, you can spend the entire year exploring all these routes one by one, but you don't have the time to just drive around the entire state of New Jersey.
And take the fastest way to get out, of course.
You have to pick the best one based on the information that you have right now.
So, how does the slime mold solve this problem?
So, when a puddle of slime mold happens to move over a food source, this is gonna trigger a set of signals inside the cell.
It's gonna start pulsating faster with a higher amplitude locally.
Now, the neighboring parts of the slime mold are gonna start, also, in turn, pulsating faster.
Pulling the cell body in the fluid away from the areas with less food.
The side that is close to the higher-priority food source little by little is going to basically redistribute all the material that is attached to the poorer food source.
When they tested this behavior in a lab using petri dish upon petri dish of slime molds, Dr. Garnier and his colleagues saw that the mold often chose the food-rich side to spread to.
In about 80% to 85% of the cases, it finds the best arm of the system.
It's a lot better than what human beings are capable of doing.
And if a single-cell blob with no brain makes better decisions than we do, maybe its behaviors are worth replicating and studying in detail.
The next step is to transform this behavior into equations and computer models.
Where they can then be applied to robots, urban planning, or maybe even solving our traffic woes.
Which is something that is actually very difficult to do experimentally by blocking an entire bridge for an entire day.