no code implementations • 6 Apr 2023 • Edric Tam, David Dunson
We propose to use the Fiedler value of the neural network's underlying graph as a tool for regularization.
1 code implementation • 28 Jan 2022 • Edric Tam, David Dunson
For comparing graphs that do not have any ambiguities due to basis symmetries (i. e. the spectrums are simple), we show that the ELD becomes a natural pseudo-metric that enjoys nice properties such as invariance under graph isomorphism.
no code implementations • ICML 2020 • Edric Tam, David Dunson
We propose to use the Fiedler value of the neural network's underlying graph as a tool for regularization.