>As a research organization, CEA has been studying how best to bring deep neural networks to edge computing. Asked about barriers to DNN on edge computing, Duranton said it’s clear that “floating point” server solutions cannot be applied, because of “power, size and latency constraints.” Other limitations include: “a number of MACs, bandwidth and on-chip memory size,” he added.
>OK. “Integer instead of floating” is one of the first things we need to consider.
I wonder why they don't try to implement Unums[1].
[1] - https://en.wikipedia.org/wiki/Unum_(number_format)