Bastian Rieck

First author of "Neural Persistence: A Complexity Measure for Deep Neural Networks Using Algebraic Topology" (with Matteo Togninalli, Christian Bock, Michael Moor, Max Horn, Thomas Gumbsch & Karsten Borgwardt; ICLR 2019), a method that scores a trained network's complexity by running persistent homology directly on its weights. Matters to the vault as the third founding-era neural-network persistent-homology diagnostic examined here — his method's specific design choice, computing over every layer's complete weighted graph with no subsampling step, turned out to share the same structural immunity to a hub-selection-style sampling artifact that the vault's other two founding NN-PH diagnostics (Birdal, Gutiérrez-Fandiño) already showed.

References

• claim-rieck-2019-neural-persistence-computed-without-subsampling
• observation-hub-selection-risk-needs-incomplete-object-nn-diagnostics-read-complete-record
• observation-hub-selection-artifact-absent-by-design-in-founding-ph-generalization-papers
• entity-tolga-birdal · entity-asier-gutierrez-fandino