The developer creates new Pokemon using the algorithm

Did you catch them? What if you invented them? Thanks to machine learning, a developer has created his own Pokémon and encourages computer enthusiasts to do the same.

Machine learning, or machine learning in French, is a branch of artificial intelligence that focuses on using data and algorithms to imitate the way humans learn, gradually improving its accuracy.

Concretely, this means that by giving access to data to certain algorithms, we can teach them to “learn”. We can thus train this kind of algorithm to generate images based on a set of data, such as the popular website This person does not exist Which generates faces of non-existent people on every new site visit.

Convert them all!

It is towards this last class of machine learning Max Wolf, who is a regular in the computer content creation business, decided to stick around. And so he set himself the challenge of getting the computer to learn what a Pokémon looks like, in order to procedurally create it itself. Thus a representation of all Pokémon was submitted to a computer that was able to learn to identify them, identify their characteristics, and then recreate them. Surprisingly, the result is quite convincing.

While some of the designs are a bit too complicated to be a real Pokemon, the majority of the Pokemon created are believable enough to fool the uninitiated. Faced with the success of his personal pocket creatures, Max Wolf decided to publish more findings.

For those more nostalgic, Max Wolf also developed an algorithm based only on 151 first-generation Pokémon.

Available on github

However, the good quality of the results obtained must be kept in mind. In fact, this type of algorithm produces a very large number of results, most of which contain significant visual errors, making them unusable. So maybe Max Wolff should have sorted out the results in order to keep only the best. Either way, the result is still very pleasant.

Those who wish can also get their hands dirty and try the algorithm for themselves. Max Wolf has already announced his algorithm github. If you wish, you can download it from this is the address.