AI brief
A question asks how high entropy soft targets relate to lower gradient variance between training cases, referencing the paper "Distilling the Knowledge in a Neural Network" by Hinton et al.
Written by AI from AI Stack Exchange's published text.
Source
“I've been trying to understand the Distilling the Knowledge in a Neural Network paper by Hinton et al.”
Interpretation or community commentary rather than straight reporting.
Read original story ↗