What is NEAT NEAT, originally developed by Kenneth Stanley and Risto Miikkulainen, stands for NeuroEvolution of Augmenting Topologies. Essentially, this is a genetic evolution model that implements an evolving topology. This means that the amount of nodes and the way they are connected evolves over time. To do this, a population of neural networks is generated. After a fitness function is used to determine the viability of each individual, high-viability individuals are used to generate a new population, which are then mutated.