Rumor has it that it’s very difficult to successfully train deep learning models, that it is “black magic” and requires years of experience to master. Ideally, given a leaning task, the model should self-adjust to the problem at hand and just work. How close we are to this goal? - Andrew Rabinovich, Principal Engineer, Magic Leap

Post a Comment

  • We reserve the right to delete any comments that we feel are disruptive.