1 repositorio
Systems designed to evolve neural network architectures and weights using genetic algorithms.
Distinct from Neural Network Training Frameworks: Focuses specifically on evolutionary training rather than gradient-based optimization frameworks
Explore 1 awesome GitHub repository matching artificial intelligence & ml · Neuroevolution Frameworks. Refine with filters or upvote what's useful.
FlappyLearning is a machine learning sandbox and neural network trainer designed to simulate neuroevolutionary strategies. It provides an environment where artificial agents are evolved using genetic algorithms to optimize performance within a game simulation. The system employs a neuroevolutionary population model that utilizes feedforward neural networks to develop autonomous game agents. It manages the evolution of these agents through generation-based iteration, using fitness-based selection to determine which networks survive and reproduce. The training framework incorporates stochastic
Provides a complete framework for evolving populations of neural networks across multiple generations.