Open kacky24 opened 5 years ago
GNNの表現力を分析する理論的な枠組みを提案し、それを元にGraph Isomorphism Network(GIN)を提案。
参考URL: https://medium.com/programming-soda/graph-neural-network%E3%81%AE%E5%87%A6%E7%90%86%E3%81%A8%E5%8A%B9%E6%9E%9C%E3%82%92%E7%90%86%E8%A7%A3%E3%81%99%E3%82%8B-how-powerful-are-graph-neural-networks-a26ee9245cce
Keyulu Xu ∗ † MIT keyulu@mit.edu Weihua Hu ∗ ‡ Stanford University weihuahu@stanford.edu Jure Leskovec Stanford University jure@cs.stanford.edu Stefanie Jegelka MIT stefje@mit.edu
https://arxiv.org/pdf/1810.00826.pdf
ICLR2019
short summary
GNNの表現力を分析する理論的な枠組みを提案し、それを元にGraph Isomorphism Network(GIN)を提案。
参考URL: https://medium.com/programming-soda/graph-neural-network%E3%81%AE%E5%87%A6%E7%90%86%E3%81%A8%E5%8A%B9%E6%9E%9C%E3%82%92%E7%90%86%E8%A7%A3%E3%81%99%E3%82%8B-how-powerful-are-graph-neural-networks-a26ee9245cce
author
Keyulu Xu ∗ † MIT keyulu@mit.edu Weihua Hu ∗ ‡ Stanford University weihuahu@stanford.edu Jure Leskovec Stanford University jure@cs.stanford.edu Stefanie Jegelka MIT stefje@mit.edu
URL
https://arxiv.org/pdf/1810.00826.pdf
year
ICLR2019