Deep Learning Breakthrough Made by Rice University Scientists

Re:Deep Learning Breakthrough Made by Rice University Scient... by allovernow on Saturday December 14, 2019 @02:45AM (#21788347)

>"[its] training times are about 7-10 times faster, and... memory footprints are 2-4 times smaller" than those of previous large-scale deep learning techniques.

Which matches the abstract. If this has general applications it's a pretty big leap to shrink model sizes and speed up training by orders of magnitude, especially at a time when many SOTA models are only feasible for well funded groups because of their size.

[ Parent | on HN ]

Re:Deep Learning Breakthrough Made by Rice University Scient... by primitivesuave on Sunday December 15, 2019 @01:06PM (#21797742)

This is not a strong result as noted by the reviewers, and they have not proven state-of-the-art performance among other things. Those of us who did ML in academia also look down upon using the media to bolster ones claims before a thorough peer review of research.

[ Parent | on HN ]