Swarm Intelligence, or distributed intelligence, is a concept based on the coordination between a massive number of individual intelligences. Find out everything you need to know about it.
In nature, many creatures are able to coordinate their individual intelligences to perform complex tasks. They demonstrate efficiency or even creativity. Take the example of ants.
A ant alone is only able to perform a limited number of actions. However, a colony of ants is able to build bridges, food and information highways. It can also go to war, or even enslave other species of ants.
This is also the case for schools of fish, clouds of birds, swarms of bees and other animal species or even bacteria. This phenomenon, which can be observed in nature, has inspired a concept that can be found in the computer field under the name of Distributed Intelligence, or Swarm Intelligence.
Swarm Intelligence or Distributed Intelligence: Definition
Swarm Intelligence or distributed intelligence is the collective behaviour of decentralized and self-organized natural or artificial systems. More specifically, it is generally a collective behaviour resulting from local interactions between several individuals or with their environment. The term was first used in 1989 by Gerardo Beni and Jing Wang in connection with cellular robotic systems.
Whether natural or artificial, distributed intelligence systems usually have several specificities. In general, they are composed of a large number of relatively homogeneous individuals whose interactions are based on simple behavioural rules exploiting only the local information that individuals exchange directly or via the environment. The overall behaviour of the system results from these interactions.
The property that truly characterizes a Swarm Intelligence system is its ability to act in a coordinated manner without the presence of an external coordinator or controller. These different properties give rise to three other characteristics that are found in distributed intelligence systems: scalability, parallelism, and resistance to breakdowns.
Scalability means that a system can maintain function while increasing in size (or its performance) without having to redefine (reprogram) the way in which the individuals who make it up interact. Parallelism means that the individuals that make up the system can perform different actions in different places at the same time.
I mean, come on, resistance to breakdowns is linked to the decentralised and self-organised nature of the system distributed intelligence systems. If one of the individuals is affected by a failure, he can easily be replaced.
Swarm Intelligence or distributed intelligence: use cases
Distributed intelligence finds many uses in modern technology. In particular, it has been used in computing and telecommunications networks since the 1990s.
In the field of health, a 1992 article already explored the possibility of using distributed intelligence to control nanobots inside the human body. to kill tumors. The U.S. military is looking to use it to… checking unmanned vehiclesNASA is developing distributed intelligence technology for planetary mapping.
This concept is also widely used in the field of robotics, so much so that we’re talking about “Swarm Robotics.” . For example, GRITS researchers have created a swarm of simple robots capable of coordinating without communicating to play the piano just by determining the position of other robots.
More recently, the Ehang airline company has created an army composed of a thousand drones capable of coordinating to create colourful shows in the sky. The University of Cambridge and Koc University are also planning to create the Energy Neutral Internet of Drones, a swarm of drones capable of sharing information or energy. In total, about 30 nations would develop swarms of intelligent drones… for military purposes.
The Blockchain, one of the most popular technologies of the moment, is also based on the concept of distributed intelligence. Indeed, the chains of blocks are comparable to colonies of ants. in how they naturally find the most efficient way to transfer large volumes of information.
Swarm Intelligence as an alternative to Big Data?
Distributed intelligence is also used in data mining. However, it can also present itself as an interesting alternative to Big Data analysis and more precisely to predictive analysis.
A company called Unaimous AI created an artificial intelligence called UNU combining the intelligence of human individuals with algorithms to make predictions. Each opinion added to the software causes “micro-changes”, allowing the user group to think collectively as a single entity. swarm-like.
This revolutionary system has been able to successfully predict the four winning Kentucky Derby horses, the Oscar winners, the NFL winners and the Stanley Cup winners. The more individuals involved in the swarm, the better the predictions.