I go on it regularly but a lot of the papers on it seem to require a lot of prior knowledge on Machine Learning. Is there any great paper's Adrian has discussed that either require no extensive prior knowledge (just great in it's own right) or serve as a good foundation for the common papers he posts?