@inproceedings{CI-StanovicGB23,


author = {Stevan Stanovic and
Benoit Ga{"{u}}z{`{e}}re and
Luc Brun},
editor = {Mario Vento and
Pasquale Foggia and
Donatello Conte and
Vincenzo Carletti},
title = {Maximal Independent Sets for Pooling in Graph Neural Networks},
booktitle = {Graph-Based Representations in Pattern Recognition - 13th {IAPR-TC-15}
International Workshop, GbRPR 2023, Vietri sul Mare, Italy, September
6-8, 2023, Proceedings},
series = {Lecture Notes in Computer Science},
volume = {14121},
pages = {113--124},
publisher = {Springer},
year = {2023},
url = {Springer:=https://link.springer.com/chapter/10.1007/978-3-031-42795-4_11, HAL:=https://hal.science/hal-04160860v1},
doi = {10.1007/978-3-031-42795-4_11},
theme = "pattern",
abstract = "Convolutional Neural Networks (CNNs) have enabled major advances in image classification through convolution and pooling. In particular, image pooling transforms a connected discrete lattice into a reduced lattice with the same connectivity and allows reduction functions to consider all pixels in an image. However, there is no pooling that satisfies these properties for graphs. In fact, traditional graph pooling methods suffer from at least one of the following drawbacks: Graph disconnection or overconnection, low decimation ratio, and deletion of large parts of graphs. In this paper, we present three pooling methods based on the notion of maximal independent sets that avoid these pitfalls. Our experimental results confirm the relevance of maximal independent set constraints for graph pooling. "

 }