In past roles, I’ve spent countless hours trying to understand why state-of-the-art models produced subpar outputs. The underlying issue here is that machine learning models don’t “think” like humans ...