An effective aggregation of node features into a graph-level representation via readout functions is an essential step in numerous learning tasks involving graph neural networks. Typically, readouts are simple and non-adaptive functions designed such that the resulting hypothesis space is permutation invariant. Prior work on deep sets indicates that such readouts might require complex node embeddings that can be difficult to learn via standard neighborhood aggregation schemes. Motivated by this, we investigate the potential of adaptive readouts given by neural networks that do not necessarily give rise to permutation invariant hypothesis spaces. We argue that in some problems such as binding affinity prediction where molecules are typically presented in a canonical form it might be possible to relax the constraints on permutation invariance of the hypothesis space and learn a more effective model of the affinity by employing an adaptive readout function. Our empirical results demonstrate the effectiveness of neural readouts on more than 40 datasets spanning different domains and graph characteristics. Moreover, we observe a consistent improvement over standard readouts (i.e., sum, max, and mean) relative to the number of neighborhood aggregation iterations and different convolutional operators.

Graph Neural Networks with Adaptive Readouts / Buterez, D.; Janet, J. P.; Kiddle, S. J.; Oglic, D.; Lio, P.. - 35:(2022). (Intervento presentato al convegno Advances in Neural Information Processing Systems (was NIPS) tenutosi a New Orleans; USA).

Graph Neural Networks with Adaptive Readouts

Lio P.
2022

Abstract

An effective aggregation of node features into a graph-level representation via readout functions is an essential step in numerous learning tasks involving graph neural networks. Typically, readouts are simple and non-adaptive functions designed such that the resulting hypothesis space is permutation invariant. Prior work on deep sets indicates that such readouts might require complex node embeddings that can be difficult to learn via standard neighborhood aggregation schemes. Motivated by this, we investigate the potential of adaptive readouts given by neural networks that do not necessarily give rise to permutation invariant hypothesis spaces. We argue that in some problems such as binding affinity prediction where molecules are typically presented in a canonical form it might be possible to relax the constraints on permutation invariance of the hypothesis space and learn a more effective model of the affinity by employing an adaptive readout function. Our empirical results demonstrate the effectiveness of neural readouts on more than 40 datasets spanning different domains and graph characteristics. Moreover, we observe a consistent improvement over standard readouts (i.e., sum, max, and mean) relative to the number of neighborhood aggregation iterations and different convolutional operators.
2022
Advances in Neural Information Processing Systems (was NIPS)
Graph Theory; Neural Network; Graph Convolutional Network
04 Pubblicazione in atti di convegno::04b Atto di convegno in volume
Graph Neural Networks with Adaptive Readouts / Buterez, D.; Janet, J. P.; Kiddle, S. J.; Oglic, D.; Lio, P.. - 35:(2022). (Intervento presentato al convegno Advances in Neural Information Processing Systems (was NIPS) tenutosi a New Orleans; USA).
File allegati a questo prodotto
File Dimensione Formato  
Buterez_Graph-neural_2022.pdf

accesso aperto

Note: https://papers.nips.cc/paper_files/paper/2022/file/7caf9d251b546bc78078b35b4a6f3b7e-Paper-Conference.pdf
Tipologia: Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 1.04 MB
Formato Adobe PDF
1.04 MB Adobe PDF

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/1727088
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 29
  • ???jsp.display-item.citation.isi??? 12
social impact