by Peter B. Nichol

Why swarm intelligence enhances business and Bitcoin

Apr 28, 2017
Artificial IntelligenceConsumer ElectronicsIT Strategy

Can intelligence be amplified by thinking together? Ants do it. Birds do it. What about humans? Collective intelligence is the next wave of intelligence. Swarm intelligence connects systems with real-time feedback loops. Individual efforts combine to form a greater value.

school of fish vortex
Credit: Thinkstock

Fish school. Birds flock. Bees swarm. A combination of real-time, biological systems blends knowledge, wisdom, opinions and intuition to unify intelligence. There’s no central control unit. These simple agents interact locally, within their environment, and new behaviors emerge.

Swarm intelligence is the self-organization of systems for collective decentralized behavior. Swarm intelligence enables groups to converge and create an independent organism that can do things that individuals can’t do on their own.

Why can’t humans swarm? Fish detect ripples in the water. Birds use motion detected through the flock. Ants leverage chemical traces. Until recently, there’s been little research conducted on “human swarming.” If nature can work together, why can’t humans use similar decision spaces to arrive at a preferred solution? Will the next generation of breakthrough innovation stem from the wisdom of the crowd — swarm intelligence?

Whether we’re talking about nature, humans or robots, swarm intelligence creates a virtual platform to enable distributed engagement from system users. Through this engagement, feedback can be provided in a closed-loop, swarming process.

Individual force for unified objectives

Swarm intelligence draws from biologically inspired algorithms to enhance robotics and mechatronics. Evolutionary optimization is more than ant-colony optimization algorithms (ACO), bee-colony optimization algorithms (BCO) or particle-swarm optimization (PSO). Swarm intelligence can be applied to immune systems, computer vision, navigation, mapping, image processing, artificial neural networks and robotic motion planning.

Bio-inspired systems bring new intelligence to the design of robotics and are used in aerial flying robots, robotic manipulators and underwater vehicles.

The physical, biological and digital worlds benefit immensely by learning from nature. These bio-inspired applications are creating swarm algorithms empowering a newly discovered digital autonomy.

Ants and distributed systems

Technology-based distributed systems are collections of independent computers that appear to work as a unified, coherent system. This same effect is found in swarms. The common element is that control is distributed across individuals or entities and communication isn’t localized.

Why is Bitcoin so fascinating to us? Could it be that the Bitcoin network is a self-organizing, collective intelligence similar to that mesmerizing school of fish?

The collective intelligence, or COIN, framework was first introduced in a paper published in 2000 by John Lawson and David Wolpert of NASA’s Ames Research Center.

This framework helped identify — using similar system attributes — where collective intelligence might exist.

  1. Multi-agent system.
  2. No central operator.
  3. No centralized communication.
  4. Unified utility function.
  5. Agents run reinforcement learning algorithms for validation.

Bitcoin is a large version of a multi-agent, reinforcement learning system. The same challenge injected into swarms is inherent in Bitcoin: How are rewards to individuals, agents or entities assigned? The social aspects of swarms are both simple and complex. Group behavior emerges as more significant than individual actions — complexity out of simplicity.

Swarms can solve more than just static problems. Units interact in localized ways and can solve online, offline, stationary, time-varying, centralized, distributed and dynamic problems.

How does a swarm live? How does a swarm communicate? A unique “life” takes shape when a swarm forms, and it has everything to do with spatial intelligence. When observing swarms, we start to notice certain principles:

  1. Work division
  2. Collective behaviors
  3. Navigation
  4. Communication
  5. Self-organization

Social survival

Dinosaurs weren’t social. Ants are social, and they have outlasted dinosaurs and are able to survive in a range of environments and climates. How do ants build their nests? How do ants navigate? Why can ants locate food fast? There’s a one-word explanation: sociality.

The key to human survival isn’t having sophisticated intelligent robots that will floss your teeth while you’re in the shower. The secret is sociality. We must build social systems when we design intelligent systems. There are many examples of nature’s social systems we can draw from:

  • An implausibility of wildebeest: They move through rivers in sheer numbers to avoid crocodiles.
  • A rabble of butterflies: Monarch butterflies migrate to escape the cold North American winters.
  • A rookery of penguins: Emperor penguins converge in a huddle to stay protected from the Antarctic winters.
  • A business of mayflies: Use swarms of 8,000 to attack predators in volume.
  • A plague of locusts: Synchronize their wing beats to make travel more efficient.
  • A shoal of fish: Silver carp leap into the air as a unit to avoid predators.
  • A pod of dolphins: Superpods of dolphins, which can exceed 1,000 individuals, form a pod for protection and hunting.
  • A flight of birds: Budgerigars, a type of parakeet, assemble to act as a unit to make decisions, fend off predator attacks and find food.
  • A cloud of bats: a social vortex of bats forms for communication and to make decisions on foraging.

Nature’s progression and technology’s evolution are amplified with social systems. The end of social abnormalities may be the introduction of swarm intelligence.

Is there a better way to build super-intelligence?

Let’s collect lessons from nature, insights from humans and the unified benefits of intelligent systems and create something smarter than ourselves. These intelligent systems — things smarter than ourselves — appear to think and act. The algorithms, robotics and systems are only a piece of the system we’ll create. Instead of creating and designing complete intelligence systems, maybe we should apply simple rules to form collections of behaviors or swarms.

These swarms could respond by connecting real-time human insights into more intelligent systems with morals, values, emotions and empathy. Swarm intelligence won’t be something you watch on a Ted Talk. Swarm intelligence is going to be a feeling that transcends nature through a collision of the digital and physical worlds.

Tomorrow’s systems will be designed with swarm intelligence and spatial judgment.