Neural Bird demonstrates two fundamental concepts in modern computer science: Artificial Neural Networks and Genetic Algorithms.
Instead of being explicitly programmed to jump at specific coordinates, the birds learn through experience. Each bird is equipped with a simple brain that processes raw sensor data—height, velocity, and distance to obstacles—and translates it into a decision.
Through a process of selective breeding and mutation, the population evolves, preserving successful "intuitions" and discarding failures.
This mirrors how machine learning models are trained in the real world: defining a fitness function, managing a population of candidates, and iteratively optimizing parameters. It teaches an intuition for high-dimensional search spaces and how complex, emergent behavior can arise from simple mathematical rules.