What can we learn from ants, bees and fish? Well, Alot!! The Expertise of the Few; The Power of the Many; The Results that Matter! While the contribution of an individual may be negligible, put together as a group they give meaningful results that matter.
Collective Intelligence (CI) is shared or group intelligence emerging from the collaboration and competition of many individual beings. It appears in consensus decision making in bacteria, insects and animals. It relates to how the simple actions of individuals can come together to produce sophisticated behaviour of the collective. Some examples you may have come across in your everyday life:
- Ants foraging in their environments
- Bees finding and exploiting nectar in flowers
- Termites creating complex mounds
Individuals in all the above cases have extremely limited cognitive capacity, yet collectively they make intelligent systems and cope with complex situations. By applying the distributed knowledge in a population or a group and the expertise of individuals, intelligent behaviour as a collective is visible. Different people look at a problem from different point of views. They will hence, point out solutions to address the problems that they see and hence the problem gets solved as a whole. One can leverage this collective intelligence to:
- Predict behaviour and future events
- Solve complex problems
- Develop improved products
- Deliver improved services
Changing our perspective from an individual to a collective being — seeing the system as one, also changes our approach to a problem. Some of the simple yet key aspects of CI can be stated as:
- Two is company, three is crowd — and crowd is good
- Ignorance is bliss, sometimes
- Encourage random encounters
- Look for patterns in signs
- Pay attention to your neighbours
- Local information can lead to global wisdom
Collective Intelligence can broadly be classified into:
- Collecting data
- Creating collaboratively
- Expertise in a crowd
- Making decisions collectively — wisdom of crowds
- Micro crowdsourcing
- Crowd Mining
Once enough data has been collected from a crowd, efficient machine learning algorithms can be used to analyse and interpret this data to extract meaningful insights from it and hence combining machine learning with CI will yield fruitful results.
One such example of a successful implementation of collective intelligence is the “Foldit” problem statement wherein citizen scientist gamers solve decades-old problem in three weeks! The scientists at the University of Washington were working on a challenging biomedical research problem. There were many failed attempts to solve the crystal structure of an enzyme related to retroviruses. They developed an online protein-folding game called “Foldit” to recruit the help of citizen scientists to produce accurate models of the protein. In “Foldit,” anyone can learn simple folding techniques and manipulate virtual molecules. Players participate from various backgrounds and locations. Within 3 weeks participants of the game were able to solve puzzle that has vexed the scientist for years as written in an article in Scientific American.
Other successful implementations of CI include:
Collective Intelligence combined with machine learning, specially deep learning go very well together. CI can provide for the enormous data typically required by deep learning algorithms. Given the distributed nature of work that is becoming more and more typical of any project and the access to internet becoming more feasible, it makes sense to restructure existing problems so that they become more accessible to the masses that can help in solving them.
