Dataflow matrix machines (by Anhinga anhinga)

Section 2.1 "The First Pillar: Meta-learning Architectures" - a very important one-page overview.

Section 2.2 "The Second Pillar: Meta-learning the learning algorithms" - a very important two-page overview.

Section 2.3 "The Third Pillar: Generating effective learning environments and training data" - technical core of the paper (pages 9-16), I am not planning to overview this core at the moment (I think Jeff Clune is actively working on this together with his group at OpenAI, and I am pondering trying to do more with the first two pillars in the near future (with some luck)).