deepmind (specifically deepmind) has applied DL to protein folding (specifically protein folding) disclaimer: I'm not defending deepmind. deepmind does oversell a lot, but gary marcus is not fairhttps://twitter.com/GaryMarcus/status/1161786769282666496 …
i am aware of the work, and also alluded to some of the limits (computation cost is likely to be massive, generalization may be poor, etc). success on one benchmark does not a practical solution make.