A blog on recent work in reinforcement learning