Tune a small feedforward neural network to control a creature. Adjust weights to make it seek food, avoid walls, and survive.
How to play: Adjust the neural network weights using the sliders below the network diagram to control the creature's behavior. The creature has 5 sensors: food direction (left/right), wall distance (left/right), and a bias input. Outputs control left/right turning and forward speed. Try the presets for good starting points, or use Hill-Climb training to optimize weights automatically. Click on the canvas to place food manually.