A Path to Sub-Second, Encrypted Skin Cancer Detection

Re:A Path to Sub-Second, Encrypted Skin Cancer Detection by jasonhansel on Thursday June 13, 2019 @02:47PM (#20177570)

> So we replaced the Max-Pooling layer with an Average-Pooling layer,

Why not use: https://en.m.wikipedia.org/wiki/Smooth_maximum

[ Parent | on HN ]

Re:A Path to Sub-Second, Encrypted Skin Cancer Detection by jvmancuso on Thursday June 13, 2019 @05:07PM (#20178583)

Thanks for the suggestion! The reason average pooling is faster than max pooling is that the former uses only addition, which is quite a bit faster than comparison in MPC. The smooth max requires both exponentiation and division, both of which are significantly more expensive than doing the comparison in ReLU or max pooling.

[ Parent | on HN ]