Fig. 9
From: Co-MLHAN: contrastive learning for multilayer heterogeneous attributed networks
Illustration of multi-view contrastive learning. For \(L^{{{{\rm NS}}}}\), the embedding of the target entity \(v_{i}\) is under the network schema view (colored in blue), while positive (+) and negative (−) samples are under the meta-path view (colored in green). By contrast, for \(L^{{{{\rm MP}}}}\), the embedding of the target entity \(v_{i}\) is under the meta-path view (colored in green), while positive (+) and negative (−) samples are under the network schema view (colored in blue)