TY - JOUR
T1 - Influence-Based Mini-Batching for Graph Neural Networks
AU - Gasteiger, Johannes
AU - Qian, Chendi
AU - Günnemann, Stephan
N1 - Publisher Copyright:
© 2022 Proceedings of Machine Learning Research. All rights reserved.
PY - 2022
Y1 - 2022
N2 - Using graph neural networks for large graphs is challenging since there is no clear way of constructing mini-batches. To solve this, previous methods have relied on sampling or graph clustering. While these approaches often lead to good training convergence, they introduce significant overhead due to expensive random data accesses and perform poorly during inference. In this work we instead focus on model behavior during inference. We theoretically model batch construction via maximizing the influence score of nodes on the outputs. This formulation leads to optimal approximation of the output when we do not have knowledge of the trained model. We call the resulting method influence-based mini-batching (IBMB). IBMB accelerates inference by up to 130x compared to previous methods that reach similar accuracy. Remarkably, with adaptive optimization and the right training schedule IBMB can also substantially accelerate training, thanks to precomputed batches and consecutive memory accesses. This results in up to 18x faster training per epoch and up to 17x faster convergence per runtime compared to previous methods.
AB - Using graph neural networks for large graphs is challenging since there is no clear way of constructing mini-batches. To solve this, previous methods have relied on sampling or graph clustering. While these approaches often lead to good training convergence, they introduce significant overhead due to expensive random data accesses and perform poorly during inference. In this work we instead focus on model behavior during inference. We theoretically model batch construction via maximizing the influence score of nodes on the outputs. This formulation leads to optimal approximation of the output when we do not have knowledge of the trained model. We call the resulting method influence-based mini-batching (IBMB). IBMB accelerates inference by up to 130x compared to previous methods that reach similar accuracy. Remarkably, with adaptive optimization and the right training schedule IBMB can also substantially accelerate training, thanks to precomputed batches and consecutive memory accesses. This results in up to 18x faster training per epoch and up to 17x faster convergence per runtime compared to previous methods.
UR - http://www.scopus.com/inward/record.url?scp=85164537531&partnerID=8YFLogxK
M3 - Conference article
AN - SCOPUS:85164537531
SN - 2640-3498
VL - 198
JO - Proceedings of Machine Learning Research
JF - Proceedings of Machine Learning Research
T2 - 1st Learning on Graphs Conference, LOG 2022
Y2 - 9 December 2022 through 12 December 2022
ER -