Corpus ID: 214802776

Geometrically Principled Connections in Graph Neural Networks

@article{Gong2020GeometricallyPC,
  title={Geometrically Principled Connections in Graph Neural Networks},
  author={Shunwang Gong and Mehdi Bahri and Michael M. Bronstein and Stefanos Zafeiriou},
  journal={ArXiv},
  year={2020},
  volume={abs/2004.02658}
}
  • Shunwang Gong, Mehdi Bahri, +1 author Stefanos Zafeiriou
  • Published 2020
  • Computer Science
  • ArXiv
  • Graph convolution operators bring the advantages of deep learning to a variety of graph and mesh processing tasks previously deemed out of reach. With their continued success comes the desire to design more powerful architectures, often by adapting existing deep learning techniques to non-Euclidean data. In this paper, we argue geometry should remain the primary driving force behind innovation in the emerging field of geometric deep learning. We relate graph neural networks to widely successful… CONTINUE READING

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 53 REFERENCES

    SplineCNN: Fast Geometric Deep Learning with Continuous B-Spline Kernels

    VIEW 11 EXCERPTS
    HIGHLY INFLUENTIAL

    Geometric Deep Learning on Graphs and Manifolds Using Mixture Model CNNs

    VIEW 5 EXCERPTS

    FeaStNet: Feature-Steered Graph Convolutions for 3D Shape Analysis

    VIEW 12 EXCERPTS
    HIGHLY INFLUENTIAL

    SpiralNet++: A Fast and Highly Efficient Mesh Convolution Operator

    VIEW 9 EXCERPTS

    Geodesic Convolutional Neural Networks on Riemannian Manifolds

    VIEW 2 EXCERPTS

    Graph Attention Networks

    VIEW 1 EXCERPT