The document discusses research into recommender systems for social learning platforms. It evaluates content-based, collaborative filtering, memory-based, model-based, and graph-based recommender approaches using datasets from MACE, OpenScout, and MovieLens. Graph-based recommender systems performed best by addressing the sparsity problem in educational domains. The implemented graph-based recommender has been integrated into the Open Discovery Space platform and will undergo a user study to evaluate satisfaction metrics like novelty and serendipity.