NeuralNetwork

NeuralNetwork is a research project focused on implementing neural network architectures from scratch in pure Python, without relying on existing machine learning frameworks. The repository features two independent core experiments. The first is a Handwritten Digit Recognition system trained on the MNIST dataset, utilizing a classic backpropagation algorithm and a Softmax activation function on the output layer. The second module is an autonomous Flappy Bird bot powered by a genetic algorithm (neuroevolution). The neural network controlling the birds learns the optimal jumping execution based on spatial input data, selection of top-performing individuals, and custom weight/bias mutation mechanisms.