[D] which small data problems pique your interest

Motivated by a recent post (can't seem to find it now; maybe from another subreddit) about how DL architecture research is gatekeep-ed by intensive computational requirements, I'd like to ask: what are your favorite small data problems? Why do you find it interesting?

data machinelearning small small data