In this paper, cognitive scientist, AI researcher and co-founder of roboust.ai Gary Marcus presents his case for a hybrid approach to AI that combines data analysis and contextual knowledge (in contrast to the data hungry techniques that currently dominate the field):
“Recent research in artificial intelligence and machine learning has largely emphasized general-purpose learning and ever-larger training sets and more and more compute.
In contrast, I propose a hybrid, knowledge-driven, reasoning-based approach, centered around cognitive models, that could provide the substrate for a richer, more robust AI than is currently possible.”Marcus, The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence
Marcus defines the present state of AI research as being fixated on producing ‘pointillistic‘ systems which can never be reliable. The goal is to create trustworthy systems (graphic by me, visualizing the main points):
The full paper is at arXiv