A lot of advances in deep learning have come from providing more data and semi-structured information, but there are a finite amount of fields for which this ca

A lot of advances in deep learning have come from providing more data and semi-structured information, but there are a finite amount of fields for which this can apply. Using game theory dynamics can move AI to transfer of domain knowledge and can create more adaptable deep learning architectures. It could also be a pathway to "stronger" forms of AI.