Description
Deep Graph Infomax (DGI), a general approach for learning node representations within graph-structured data in an unsupervised manner. DGI relies on maximizing mutual information between patch representations and corresponding high-level summaries of graphs—both derived using established graph convolutional network architectures. The learnt patch representations summarize subgraphs centered around nodes of interest, and can thus be reused for downstream node-wise learning tasks. In contrast to most prior approaches to unsupervised learning with GCNs, DGI does not rely on random walk objectives, and is readily applicable to both transductive and inductive learning setups.
Description and image from: DEEP GRAPH INFOMAX
Papers Using This Method
Graph Pre-Training Models Are Strong Anomaly Detectors2024-10-24Dynamic Gradient Influencing for Viral Marketing Using Graph Neural Networks2024-03-19On the Adversarial Robustness of Graph Contrastive Learning Methods2023-11-29DGI: Easy and Efficient Inference for GNNs2022-11-28Models and Benchmarks for Representation Learning of Partially Observed Subgraphs2022-09-01Rethinking and Scaling Up Graph Contrastive Learning: An Extremely Efficient Approach with Group Discrimination2022-06-03Contrastive Laplacian Eigenmaps2022-01-14HDMI: High-order Deep Multiplex Infomax2021-02-15SCE: Scalable Network Embedding from Sparsest Cut2020-06-30Deep Graph Infomax2018-09-27