A new primitive, ExactLine, partitions any line in the input space of a piecewise-linear neural network into segments where the network is affine, enabling exact decision-boundary analysis, exact integrated gradients, and new empirical tests of adversarial-example theory.
Goodfellow, Jonathon Shlens, and Christian Szegedy
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
-
Computing Linear Restrictions of Neural Networks
A new primitive, ExactLine, partitions any line in the input space of a piecewise-linear neural network into segments where the network is affine, enabling exact decision-boundary analysis, exact integrated gradients, and new empirical tests of adversarial-example theory.