Jan Pichowski presents 'MDL-Pool: Adaptive Multilevel Graph Pooling Based on Minimum Description Length' in IceLab October 1st 11.00, in IceLab and in Zoom.
Graph pooling compresses graphs and summarises their topological properties and features in a vectorial representation.
It is an essential part of deep graph representation learning and is indispensable in graph-level tasks like classification or regression.
Current approaches disregard the interdependencies between structures at different hierarchical levels and do not adapt to datasets that contain graphs with different sizes that may require pooling with various depths.
To address these issues, we propose MDL-Pool, a pooling operator based on the minimum description length (MDL) principle, whose loss formulation explicitly models the interdependencies between different hierarchical levels and facilitates a direct comparison between multiple pooling alternatives with different depths.
Drop in to IceLab (find using mazemap) or join via Zoom: IceLab Living Room Zoom 612 568 7586
Guest researcher Jan von Pichowski is a PhD candidate at the Universität Würzburg in Germany. He is member of the Machine Learning for Complex Networks group.
Jan is a guest researcher in IceLab visiting Professor Martin Rosvall. He is also taking part in IceLab Camp September 23-26.
Jan von Pichowski, PhD candidate at the Universität Würzburg in Germany