Pure max-plus/min-plus deep networks are shown to be the opposite of universal approximators, while inserting O(N)-parameter linear scalings per layer yields provably universal morphological networks.
Given this, the ideal weight initialization follows the same approach as DEP networks withλ= 1/2
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Training Deep Morphological Neural Networks as Universal Approximators
Pure max-plus/min-plus deep networks are shown to be the opposite of universal approximators, while inserting O(N)-parameter linear scalings per layer yields provably universal morphological networks.