I can't find the journal article, but I saw an academic machine learning paper that claimed our whole lecture system was inefficient and not optimal. Our lectur

I can't find the journal article, but I saw an academic machine learning paper that claimed our whole lecture system was inefficient and not optimal. Our lectures are currently based off an old system, before all the technology now. The game changed when we introduced computers/phones/tablets, but our method didn't change. Students don't have much to rely on to determine how well they're doing in a class. Currently, we rely on practice problems. Often times, this feedback loop isn't closed. If it is, it's possible to learn where you went wrong, but often times the solution manuals say "as an exercise for the reader..." or have similar lines. This paper was proposing to devise some test to determine how well a student understood the material and suggest strategies to more easily solve the problems (afaik at least). Something I've noticed: we still rely on someone else relaying the information to us. Who says there's who explains the concept better somewhere else? Can we devise a method to have the student give their full and undivided attention while incorporating the technology?