In this project we have dealt with the reconstruction of regulatory networks using machine learning. We have reconstructed networks in unsupervised as well as supervised way. This repository consists of the scripts associated with the project. For unsupervised learning, we have used methods like lasso with stability selection and tree based models like Random Forests, Extra Trees and Gradient Boosting. For supervised learning, we have used methods from the literature concerning graph based learning like local classification models and matrix factorization to integrate global knowledge. We also tried out deep learning architectures mainly denoising autoencoders to learn the latent state of the cell.