Collective intelligence: goodbye to driving schools
What do collective or swarm intelligence and driving school teachers have in common? It may not be a direct relationship but it unleashes all its power for several reasons.
“Swarm” and “intelligence”. Remember these two words together, because you won’t stop reading and hearing about them soon. In other words “swarm intelligence” could be deemed as “collective intelligence” or “swarm behaviour”. It’s a term that will help us to explain how the behaviour of a swarm of bees or a colony of ants (yes, you read that right, the kinds of bees and ants we see in the countryside) can accelerate the arrival of self-driving cars.
We already knew that nature is smart. But the possibility of also learning from it is something we haven’t made much progress with. Researchers have begun to realize that they can use swarm intelligence in many areas of society. Transforming different fields such as robotics, data mining and medicine could produce incredible results. It’s an idea that is undoubtedly disruptive, revolutionary and extremely ambitious.
But what is it that makes collective or swarm intelligence really interesting? The answer lies in the terms “collective” and “intelligence”: a group of simple creatures, following simple rules for a complex purpose. Easy, right? Swarm intelligence is the concept of valuing the work of each individual to achieve and improve a collective goal in the most optimal way possible. There’s nothing else behind the above definition. Applying this simple equation may lead to a surprising number of effective and creative solutions.
The concept was introduced by Gerardo Beni and Wang Jing in 1989 and it focuses on the study of the collective behaviour of decentralized, self-organized, natural and artificial systems. These are made up of a set of free agents that tend to interact with each other through a medium in order to obtain a response to a problem. Inspired by nature, we can find several examples in the organization of ant colonies, in the alignment of birds in flight and in the behaviour of herds while they graze.
From theory to practice
Let’s apply the theory: we’ll observe, for example, a colony of ants. Each of them can only perform a limited range of functions, but, in a group, thanks to their collective intelligence and their communication via their pheromones, they can find food, overcome obstacles and even enslave other species of ants. The most amazing thing of all is that none of the individual ants has full knowledge of what they’re doing collectively. As a result of the algorithms inspired by these insects, we can now develop highly powerful practical applications, such as the improvement of packet routing in network systems, the construction of analytical models by means of machine learning and even, at a more abstract level, diagnoses of lung cancer with greater efficiency than any alternative or classical methods.
Can all the above be applied to self-driving vehicles? From a bird’s eye view, cars look rather like small ants wandering along a pre-established route. The big difference lies in the communication between them. And this is what gives us the key to the proper functioning of self-driving cars. If we consider that each car is a free agent and that these free agents are connected to each other and that we have billions of them, could we achieve a functional and collective kind of behaviour to reduce traffic jams, accidents and even delays on the roads? The answer can be obtained by observing nature itself.
The principle of action-reaction and the so-called boomerang effect prove it: every cause has an effect and every effect becomes the cause of something else. And this happens in today’s world with self-driving cars, swarm intelligence and driving school teachers. If it isn’t going to be necessary to obtain a driving licence, why will we need driving school teachers?
Collective or swarm intelligence is nothing more than the coming together of many isolated but connected minds thinking for a single abstract mind. So, if you’re thinking about your next job, you’d be better off learning how to program self-driving cars, because that will be as close as you get to teaching someone to drive.