Evolution is not just something we have learnt often using fossils or from the past. Sometimes it happens in real time & in real world. Computational technology with digital evolution system & field observation can solve novel evolutionary methods.And now we can observed organisms from bacterias to human in real time which can adapt their behaviour in response to environment artificially. This simply means that we can observe evolution with our eyes in the real time.
At every successive generation, the computers can randomly mutated different animal models to create new & transgenic organisms. Innovative simulation software then evaluates the models to survive more leads to evolution. It works like natural selection which mimics the evolutionary process.
Neuroevolution is a form of machine learning adapt to modify artificial neural networks & topologies that uses genetic evolutionary algorithms. In short it is used to evolve an artificial brain. Evolutionary process imitate the genetic algorithm through: Fitness evaluation, Descent with modification & selection. In the process, each artificial brains are tested on the basis of performed task. Brain which perform well are selected to reproduce into next generation. The outcome of that generation are created as modified copies of their parents with some genetic modification.
Researchers at UCSC evolved an antenna for one of NASA’s satellites with the application of genetic algorithm which performed better than the antenna designed by the engineers. The advantage of neuroevolution is bringing development of artificial intelligence with evolution of human level intelligence.
The post Scifi of Neuroevolution appeared first on Science Blog.