Disentangled Graph Convolutional Networks (Paper Explained) скачать в хорошем качестве

Disentangled Graph Convolutional Networks (Paper Explained)

TLDR: Learn disentangled node representation by automatically inferring the latent factor between edges in a graph. http://proceedings.mlr.press/v97/ma19... *Authors*: Jianxin Ma, Peng Cui, Kun Kuang, Xin Wang, Wenwu Zhu *Abstract*: The formation of a real-world graph typically arises from the highly complex interaction of many latent factors. The existing deep learning methods for graph-structured data neglect the entanglement of the latent factors, rendering the learned representations non-robust and hardly explainable. However, learning representations that disentangle the latent factors poses great challenges and remains largely unexplored in the literature of graph neural networks. In this paper, we introduce the disentangled graph convolutional network (DisenGCN) to learn disentangled node representations. In particular, we propose a novel neighborhood routing mechanism, which is capable of dynamically identifying the latent factor that may have caused the edge between a node and one of its neighbors, and accordingly assign- ing the neighbor to a channel that extracts and convolutes features specific to that factor. We theoretically prove the convergence properties of the routing mechanism. Empirical results show that our proposed model can achieve significant performance gains, especially when the data demonstrate the existence of many entangled factors. 00:00 - Introduction & Algorithm overview 08:10 - Iterative routing mechanism 16:10 - Hypothesis behind the learning algorithm 18:24 - Learning algorithm in details 28:05 - Reiterate algorithm end-to-end 31:00 - Experiments 36:24 - Conclusion Connect with me:   / andreimargeloiu