Skip to content

Latest commit

 

History

History
 
 

demos

README.md

Table of Contents

This folder contains several examples of applying graph machine learning (ML) algorithms on network-structured data to solve several common problems including node attribute inference (inferring entity properties) and link prediction (predicting relations and/or their properties). The examples demonstrate using the StellarGraph library to build machine learning workflows on both homogeneous and heterogeneous networks.

Each folder contains one or more examples of using the StellarGraph implementations of the state-of-the-art algorithms, attri2vec [4], GraphSAGE [3], HinSAGE, GCN [6], GAT [7], Cluster-GCN [10], PPNP/APPNP [9], Node2Vec [1], and Metapath2Vec [2]. GraphSAGE, HinSAGE, and GAT are variants of Graph Convolutional Neural networks [6]. Node2Vec and Metapath2Vec are methods based on graph random walks and representation learning using the Word2Vec [5] algorithm. attri2vec[4] is also based on graph random walks, and learns node representations by performing a mapping on node attributes.

The examples folder structure is shown below.

  • /embeddings

    Examples of unsupervised node representation learning for homogeneous networks, heterogeneous networks, and homogeneous networks with node features using Node2Vec, Metapath2Vec, and Unsupervised GraphSAGE algorithm, respectively.

  • /link-prediction

    Examples of using StellarGraph algorithms for link prediction on homogeneous and heterogeneous networks.

  • /node-classification

    Examples of using StellarGraph algorithms for node classification on homogeneous and heterogenous networks.

  • /ensembles

    Examples of using ensembles of graph convolutional neural networks, e.g., GraphSAGE, GCN, HinSAGE, etc., for node classification and link prediction. Model ensembles usually yield better predictions than single models, while also providing estimates of prediction uncertainty as a bonus.

  • /calibration

    Examples of calibrating graph convolutional neural networks, e.g., GraphSAGE, for binary and multi-class classification problems.

  • /community_detection

    Examples of using unsupervised GraphSAGE embeddings in a context of community detection. Community detection is demonstrated on a terrorist network, where groups of terrorist groups are found using dbscan on top of the graphSAGE embeddings. Note that this demo requires the installation of igraph-python, see the README.md in this directory for more details.

  • /interpretability

    Examples of using saliency map based methods, such as integrated gradients [11], to provide interpretability to the graph neural networks, e.g., GCN. Saliency maps are used to approximate the importance of the nodes and links (in the ego network of a target node) while making the prediction.

  • /use-cases

    Example use-cases/applications for graph neural network algorithms.

References

  1. Node2Vec: Scalable Feature Learning for Networks. A. Grover, J. Leskovec. ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2016. (link)

  2. Metapath2Vec: Scalable Representation Learning for Heterogeneous Networks. Yuxiao Dong, Nitesh V. Chawla, and Ananthram Swami. ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 135–144, 2017 (link)

  3. Inductive Representation Learning on Large Graphs. W.L. Hamilton, R. Ying, and J. Leskovec arXiv:1706.02216 [cs.SI], 2017. (link)

  4. Attributed Network Embedding via Subspace Discovery. D. Zhang, Y. Jie, X. Zhu and C. Zhang, arXiv:1901.04095, [cs.SI], 2019. (link)

  5. Distributed representations of words and phrases and their compositionality. T. Mikolov, I. Sutskever, K. Chen, G. S. Corrado, and J. Dean. In Advances in Neural Information Processing Systems (NIPS), pp. 3111-3119, 2013. (link)

  6. Semi-Supervised Classification with Graph Convolutional Networks. T. Kipf, M. Welling. ICLR 2017. arXiv:1609.02907 (link)

  7. Graph Attention Networks. P. Velickovic et al. ICLR 2018 (link)

  8. On Calibration of Modern Neural Networks. C. Guo, G. Pleiss, Y. Sun, and K. Q. Weinberger. ICML 2017. (link)

  9. Predict then propagate: Graph neural networks meet personalized PageRank. J. Klicpera, A. Bojchevski, A., and S. Günnemann, ICLR, 2019, arXiv:1810.05997.(link)

  10. Cluster-GCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networks. W. Chiang, X. Liu, S. Si, Y. Li, S. Bengio, and C. Hsiej, KDD, 2019, arXiv:1905.07953.(link)

  11. Axiomatic Attribution for Deep Networks. Mukund Sundararajan, Ankur Taly and Qiqi Yan. ICML 2017. (link)