Integrating scikit-learn into Syft

Openmined's Syft library provides an infrastructure for computing on data you do not own and cannot see. It allows data scientists to work with data owners and provide machine learning solutions while preserving the privacy of the data. Currently, its infrastructure allows users to train models using libraries like PyTorch. While Pytorch implements many deep learning models, it doesn't implement many machine learning algorithms like SVM and Random forest. This limits users of Syft in terms of using variety of algorithms to train models on remote data. Scikit-learn provides many simple and efficient tools for predictive data analysis and is very popular and is used vastly in the Data-Scientist community. This project aims to integrate Scikit-learn into Syft's infrastructure to allow users of Syft to use all end-to-end algorithms and features which Scikit-learn provides.