Lossless graph condensation is achieved through expanding window matching, outperforming existing methods.
The author proposes GEOM, a method for graph condensation via expanding window matching, to achieve lossless graph condensation by bridging previously neglected supervision signals and optimizing the condensed graph effectively.