What is Swarm Intelligence and How it works?
Social creatures working as a unified entity can perform much better in terms of solving a problem or making a decision as compared to a vast majority of the entity working individually. Many minds are better than one.
The above image shows a flock of birds moving together. You might have seen bees and fishes doing something similar. In general, these groups are collectively called swarm.
As defined on Wikipedia — “Swarm intelligence (SI) is the collective behaviour of decentralized, self-organized systems, natural or artificial.” Examples of swarm intelligence in natural systems include ant colonies, bee colonies, bird flocking, hawks hunting, animal herding, bacterial growth, fish schooling and microbial intelligence. The below TEDx talk video describes a swarm as brain of brains.
Swarm Intelligence has many use cases.
Some of them are listed below:
- Internet of Things: Powerful decentralized algorithms supported by SI are used to logically control operations of complex IoT systems. By applying swarm intelligence algorithms for IoT devices, we can provide major advantages for energy saving in IoT devices.
- Human Swarms: Humans don’t have the natural ability to form a Swarm Intelligence. Forming such intelligence requires tight feedback loops among members. Some companies like unanimous.ai are leveraging high-speed networking to form real-time systems that enable “human swarms” to converge online, combining the knowledge, wisdom, insights, and intuitions of diverse groups into a single emergent intelligence.
- Healthcare: Researchers have used ‘swarm learning’ to detect blood cancer, lung diseases and COVID-19 in data stored in a decentralized fashion. ASCA (Ant System-based Clustering Algorithm) has been used in image segmentation producing better results than 1D-SOM, k-Means, FCM and PFCM algorithms in the detection of small, atypical regions of the image.
- Pervasive Computing: When the computational capability is embedded into everyday objects such that they can effectively communicate and perform useful tasks without end-user interaction, it’s a form of pervasive computing. Distributed perceptive networks which are used to implement pervasive computing utilize Swarm Intelligence Algorithms. Multi-hop networking, optical network optimizations, and insect drones are some examples.
- Mobile Ad hoc NETwork (MANET): MANET is a decentralized network with the ability to scale to thousands of connections. They self-learn and organize as they operate and adapt as new and old nodes enter and exit the network under dynamic conditions. Swarm Autonomous Routing Algorithm (SARA) protocol can manage node to node communications in a MANET. SI can also be used to implement a distributed location service in a MANET
- Data Science & ML: By applying swarm intelligence in parallel we can achieve better results in Hydrological Forecasting. SI can optimize the parameters of an artificial neural network. SI has many other applications in data science and machine learning.
- Swarm Robotics: Swarm robotics is a field of multi-robotics in which many robots can coordinate in a distributed and decentralized way.
How Swarm Intelligence Works?
In general, swarm systems work in the following way :
- There are be a large number of distributed agents who work/act in-parallel
- The agents transmit/ receive signals from the other agents. It is through these signals that all interactions happen.
- Each agent individually processes information. There is no central command. In that sense, the agents are autonomous. But the agents are strongly influenced by what others in the system do.
- As the interaction evolves, a swarm intelligence emerges, and the group’s behaviour self organizes. The entire group decides on something that each agent might not have been able to do.
- If some of the agents are destroyed, others will quickly adapt and the system will continue to function as earlier.
In general, the following principles can be used to describe behaviour that leads to swarm intelligence
- Proximity Principle: The basic units of a swarm should be capable of simple computation related to its surrounding environment.
- Quality Principle: A swarm should be able to respond to quality factors such as determining the safety of a location.
- Principle of diverse response: The distribution is designed so that each agent will be maximally protected facing environmental fluctuations.
- Principle of stability: The group should not change their mode of behaviour every time the environment changes
- Principle of adaptability: The swarm is sensitive to the changes in the environment that result in different swarm behaviour
References & Further Reading
- Swarm Intelligence — Wikipedia
- What is Swarm AI Technology — Unanimous.ai
- Swarm Intelligence — Science Direct
- Swarm Intelligence — Research Gate, Queen’s University, School of Computing Technical Reports
- Swarm Intelligence — Tech Ferry
- Building an Identification Model Using Swarm Intelligence and Its Applications
- Swarm Intelligence — Singularity Group Blog
- The biological principles of swarm intelligence
- Swarm Intelligence — Yichen Hu
About the Author
Arpit is a seasoned technologist with vast experience in leading large cross-functional and cross-geography teams. Arpit also consults clients on competitive market analysis, defining MVPs, product ideation, product monetisation and go live strategies.
Arpit believes we should all contribute back to society. He has set his goals for social work in five broad areas. You can read more about the same in his blog post “Do Good, Together” on Tumblr. Arpit is interested in working with people who want to contribute toward the same goals.
You can follow Arpit on Linkedin and Twitter
ABC. Always be clappin’.
