Incremental GNN Embedding Computation on Streaming Graphs

Qiange Wang 1 , Haoran Lv 2 , Yanfeng Zhang 2 , Weng-Fai Wong 1 , Bingsheng He 1 · 2026

Read Incremental GNN Embedding Computation on Streaming Graphs by Qiange Wang 1 , Haoran Lv 2 , Yanfeng Zhang 2 , Weng-Fai Wong 1 , Bingsheng He 1 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-20622.txt.

Description

Wang and co-authors speed up graph neural networks on graphs that change over time. Fresh embeddings must track each new change, which costs heavy multi hop walks. Much of the graph stays the same between steps, though. Their method splits the work into fine operators that can be reordered with care. Only the touched subgraph then needs fresh math. A GPU and CPU split system holds old results at scale. Tests cut total work by 64 to 99 percent.

Catalog record

CREATORQiange Wang 1 , Haoran Lv 2 , Yanfeng Zhang 2 , Weng-Fai Wong 1 , Bingsheng He 1
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE73073
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.20622

Browse the full library · About the Tower of Babel