A trained MLP can be rewritten as a sparse additive model of one- and two-variable effects, called a Partial Response Network, with no loss in classification accuracy on benchmark tabular data.
(10) This uses the fact that when the MLP has converged, the cost function is at an extremum hence the first derivative vanishes
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
The Partial Response Network: a neural network nomogram
A trained MLP can be rewritten as a sparse additive model of one- and two-variable effects, called a Partial Response Network, with no loss in classification accuracy on benchmark tabular data.