A stable rank network with a known Lipschitz constant certifies robustness radii for persistence diagrams, and on ORBIT5K it keeps high robust accuracy where a standard PersLay model collapses.
(2020): each point in the persistence diagram is upsampled from 2 dimensions to 25 13 Agerberg Guidolin Martinelli Hoefgeest Eklund Scolamiero through 2 layers
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Certifying Robustness via Topological Representations
A stable rank network with a known Lipschitz constant certifies robustness radii for persistence diagrams, and on ORBIT5K it keeps high robust accuracy where a standard PersLay model collapses.