1 paper
Itsuki Nakagawa, Kenji Yamanishi
We propose an information-theoretic framework for graph novelty generation, which aims to generate data that are distinct from existing patterns while preserving global structural…