Abstract: Split learning (SL) is a privacy-focused method for training deep learning models that prevents clients and servers from sharing raw data. There are a few downsides to supervised learning, ...