“Scientists of AI at Google’s Google Brain and DeepMind units acknowledge ML is falling short of human cognition and propose that using models of networks might be a way to find relations between things that allow computers to generalize more broadly about the world.” Peter W. Battaglia of Google’s Deep Mind unit, along with colleagues from Google Brain, MIT, and the University of Edinburgh argue for “blending powerful deep learning approaches with structured representations,” and their solution is something called a “graph network.”

These are models of collections of objects, or entities, whose relationships are explicitly mapped out as “edges” connecting the objects. The idea is that graph networks are bigger than any one ML approach. Graphs bring an ability to generalize about structure that the individual neural nets don’t have.

Once approved for inclusion, rules could be used to identify issues with missing or incorrect data. For example, to flag as an issue a blue eyed person whose parents have brown eyes. Established rules can also be used as “known axioms” in further application of ML to ensure that insufficient or incorrect training data does not cause erroneous conclusions.