|World cup / Photo by: SPF via Shutterstock|
San Francisco-based technology firm Unanimous A.I.predicts Germany will beat Brazil and will emerge as the champion of the 2018 World Cup. It also claims Spain and France will vie for third and fourth places, according to Luke Dormehl, writing for Digital Trends.
To come up with such a forecast, Unanimous A.I. relied on swarm artificial intelligence technology that integrates human insights with AI algorithms. Such a union makes for a system that is better than just either humans or computers only. The technology is patterned after swarms in nature - bees, fishes or birds - that use the insights of large groups.
Unanimous A.I.’s swarm technology relies on similar principles to resolve complex issues, such as providing probability-based outcomes for every match in the World Cup. Company CEO Louis Rosenberg said swarm technology can be used in solving problems without undergoing training on specific data sets. Rosenberg said they are tapping the wisdom, experience, and insights of people who know soccer a lot. Their algorithms have already been trained in human behaviors.
It has enlisted the help of British firm Colossus which recruited a team of 30 sports handicappers and put them at Unanimous A.I.’s disposal. The handicappers were connected together through the Internet and used Unanimous A.I.’s algorithms to prophesize who will be victorious in each of the World Cup Games. The task took more than an hour to finish but they were able to come up with a complete World Cup bracket.
On the other hand, German and Belgian scientists are using a machine learning method called random forest to predict the World Cup winners. A random forest is a group of decision trees that predict future events on a step-by-step basis, in which the result at each branch is governed by a particular data set. It attempts to overcome the shortcomings of traditional decision tree methods by studying random branches numerous times at the same time. The final result is basically the average of a large number of trees built through the random forest method.