search

Shen-Lab / GraphCL_Automated

[ICML 2021] "Graph Contrastive Learning Automated" by Yuning You, Tianlong Chen, Yang Shen, Zhangyang Wang; [WSDM 2022] "Bringing Your Own View: Graph Contrastive Learning without Prefabricated Data Augmentations" by Yuning You, Tianlong Chen, Zhangyang Wang, Yang Shen
108 stars 11 forks source link

About the min-max modeling. #3

Open cjfcsjt opened 2 years ago

cjfcsjt commented 2 years ago

The research work is exciting! I noticed that you formulate the automated graph contrastive learning as the min-max optimization problem. May I ask that why not use the bi-level optimization where the upper level objective is to search the augmentation operations and the lower level objective is to optimize the contrastive loss (like existing one-shot neural architecture search methods)?

yyou1996 commented 2 years ago

Hi @cjfcsjt,

Sorry for the late reply. In paper we did solve it by a naive bi-level optimization if referring to section 3.2.