Deep Learning vs. Evolutionary Computation

Nowadays, nobody can deny the power of Machine Learning and its most successful approach: Deep Learning. As a matter of fact, with Deep Learning, we now can do so many incredible things that were very hard just several years ago. From security control, healthcare, finance, to autonomous vehicles and robotics … Deep Learning has been emerging as one of the most powerful technologies that are transforming the world today.

However, talking so much about Deep Learning seems to make many people lose focus on another important branch of AI: Evolutionary Computation!

If you have read a previous post: Branches of Artificial Intelligence, you may remember that along with Machine Learning and Fuzzy Logic, Evolutionary Computation is an important branch of Soft Computing (or Computational Intelligence).

Indeed, Evolutionary Computation focuses on optimization algorithms that are inspired by the Biological Evolution. Its general idea is very natural: starting from an initial set of candidate solutions called population, new generations are iteratively created by removing less desired solutions and randomly introducing small changes to the population. The process continues until finding an acceptable solution to the problem. By this idea, Evolutionary Computation allows an exploratory search that continuously refines a set of candidate solutions, therefore it can discover novel ideas and creative solutions that may be surprising!

Evolutionary Computation has been successfully applied in many real-life applications and research fields such as Evolvable hardware, a field that focuses on creating special hardware that can dynamically change its structure and behaviors by interacting with its environment. Evolutionary Computation can also be used in Gaming in order to create and evolve game characters. Or it has been used to solve planning problems such as the Vehicle Routing Problem or the Traveling Salesman Problem. Another well-known application of Evolutionary Computation is to design antenna’s shapes to create the best radio patterns such as the 2006 NASA ST5 spacecraft antenna.

Evolved antenna - Wikipedia

In comparison with Deep Learning, Evolutionary Computation is more creative as it focuses on searching undiscovered solutions while Deep Learning focuses on modeling knowledge from data. There is no doubt that Deep Learning is doing many incredible things and has proven its value, but Evolutionary Computation will be the next step forward from Deep Learning to bring us to the AI of the future when intelligent agents will have the ability to « think » creatively – it’s is the CREATIVE AI.

Check this AI comprehensive course that covers Machine Learning, Evolutionary Computation, and Fuzzy Logic from scratch!

Thanks for reading & sharing.

Votre commentaire

Entrez vos coordonnées ci-dessous ou cliquez sur une icône pour vous connecter:

Logo WordPress.com

Vous commentez à l’aide de votre compte WordPress.com. Déconnexion /  Changer )

Photo Google

Vous commentez à l’aide de votre compte Google. Déconnexion /  Changer )

Image Twitter

Vous commentez à l’aide de votre compte Twitter. Déconnexion /  Changer )

Photo Facebook

Vous commentez à l’aide de votre compte Facebook. Déconnexion /  Changer )

Connexion à %s