Model-Based Methods in Reinforcement Learning - Part 4: Beyond Vanilla MBRL