An unknown group of researchers has developed a new approach to artificial intelligence (AI) that solves real-world problems. Instead of developing AI in a centralised, top-down manner, the researchers have embraced the bottom-up approach that mimics the way nature solves problems. This approach, known as NeuroEvolution of Augmenting Topologies (NEAT), uses a process similar to natural evolution to create AI.
NEAT begins with a population of simple neural networks, which are then tested on a problem. The best-performing networks are selected and ‘bred’ to create a new generation of networks. Over time, the networks evolve, becoming increasingly complex and better at solving the problem. This approach has been successful in a range of tasks, from controlling robots to playing video games.
The researchers believe that NEAT could be used to solve real-world problems, such as optimising traffic flow or managing energy grids, by evolving AI to find the best solution. They argue that this approach is more democratic and accessible than traditional AI development, as anyone can use it to create AI for their own needs.
Despite its potential, NEAT has not been widely adopted in the AI community. The researchers hope to change this by making their work more visible and demonstrating the value of their approach. They believe that by giving power to the people, AI can be used to solve real-world problems in a democratic and accessible way.
Go to source article: https://medium.com/@atduskgreg/power-to-the-people-how-one-unknown-group-of-researchers-holds-the-key-to-using-ai-to-solve-real-cc9e75b1f334#.maviae4xx