tladeras’s avatartladeras’s Twitter Archive—№ 4,994

    1. …in reply to @heatherklus
      @heatherklus This. Every data type is different and the mark of a good ML engineer is their ability to adapt their approaches to the quirks of the new data types they encounter.
  1. …in reply to @tladeras
    @heatherklus Red flags for me: hearing people say things like "I don't really care to know more about the data, I *know* this approach will work." I realize that there are useful abstractions, but integrating priors and understanding the data will still get you farther than a one-fits-all.