HOPPER: Learnable Hop Extraction for Linearized Graph Sequence Models
HOPPER is introduced, an end-to-end learnable extension of LGSM that learns how hop sequences should be extracted before processing by a modern state-space model, and demonstrates that learnable sequence extraction is a flexible and effective framework for long-range graph representation learning across synthetic, phys...
Isuru Herath, Arin Gopakumar, Sharan Sahu
· 0 citations