Presheaf and copresheaf models of continuous functions on spaces provide a functorial view of neighborhood-aggregating networks and obstructions that explain some of their empirical shortcomings.
Multilayer feedforward networks are universal approxi- mators
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Functorial Formulation of Neighborhood Aggregating Deep Learning
Presheaf and copresheaf models of continuous functions on spaces provide a functorial view of neighborhood-aggregating networks and obstructions that explain some of their empirical shortcomings.