Balancing performance and complexity with adaptive graph coarsening

0585922 - ÚI 2025 C - Conference Paper (international conference)
Dědič, M. - Bajer, L. - Procházka, P. - Holeňa, Martin
Balancing performance and complexity with adaptive graph coarsening.
The Second Tiny Papers Track at ICLR 2024. OpenReview.net / ICLR, 2024.
[ICLR 2024. International Conference on Learning Representations /12./. Vienna (AT), 07.05.2024-11.05.2024]
Institutional support: RVO:67985807
Keywords : Graph representation learning * Graph coarsening * Performance-complexity trade-off * HARP
OECD category: Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
https://openreview.net/forum?id=DrHwIzz93C

We present a method for graph node classification that allows a user to precisely select the resolution at which the graph in question should be simplified and through this provides a way of choosing a suitable point in the performance-complexity trade-off. The method is based on refining a reduced graph in a targeted way following the node classification confidence for particular nodes.
Permanent Link: https://hdl.handle.net/11104/0353559