推荐一个GitHub，里面收集整理了很多深度图神经网络Deep GNN的研究工作，这也是近两年图神经网络的研究热点之一。
链接：github.com/mengliu1998… 为什么要专门研究Deep GNN呢？这是由于GNN通常在1-2层效果较好，随着层数的增加，GNN的表现会大幅度下降。传统DNN中也有这个问题，Kaiming He的ResNet就是一个很著名的解法。 尽管这两年关于GNN的深度问题有各种研究和解释，比如过平滑，但是GNN深层退化现象是不是仅仅由于过拟合呢？比如，19ICLR PPNP这篇就提到了过拟合是Deep GNN退化的原因之一。 2021
[arXiv 2021] Two Sides of the Same Coin: Heterophily and Oversmoothing in Graph Convolutional Neural Networks arxiv.org/abs/2102.06… [arXiv 2021] Graph Neural Networks Inspired by Classical Iterative Algorithms arxiv.org/abs/2103.06… [ICML 2021] Training Graph Neural Networks with 1000 Layers arxiv.org/abs/2106.07… github.com/lightaime/d… [ICML 2021] Directional Graph Networks arxiv.org/abs/2010.02… github.com/Saro00/DGN [ICLR 2021] On the Bottleneck of Graph Neural Networks and its Practical Implications openreview.net/forum?id=i8… github.com/tech-srl/bo…) [ICLR 2021] Adaptive Universal Generalized PageRank Graph Neural Network openreview.net/forum?id=n6… github.com/jianhao2016… [ICLR 2021] Simple Spectral Graph Convolution openreview.net/forum?id=CY… 2020
[arXiv 2020] Deep Graph Neural Networks with Shallow Subgraph Samplers arxiv.org/abs/2012.01… [arXiv 2020] Revisiting Graph Convolutional Network on Semi-Supervised Node Classification from an Optimization Perspective arxiv.org/abs/2009.11… [arXiv 2020] Tackling Over-Smoothing for General Graph Convolutional Networks arxiv.org/abs/2008.09… [arXiv 2020] DeeperGCN: All You Need to Train Deeper GCNs arxiv.org/abs/2006.07… github.com/lightaime/d… [arXiv 2020] Effective Training Strategies for Deep Graph Neural Networks arxiv.org/abs/2006.07… github.com/miafei/Node… [arXiv 2020] Revisiting Over-smoothing in Deep GCNs arxiv.org/abs/2003.13… [NeurIPS 2020] Graph Random Neural Networks for Semi-Supervised Learning on Graphs proceedings.neurips.cc/paper/2020/… github.com/THUDM/GRAND [NeurIPS 2020] Scattering GCN: Overcoming Oversmoothness in Graph Convolutional Networks proceedings.neurips.cc/paper/2020/… github.com/dms-net/sca… [NeurIPS 2020] Optimization and Generalization Analysis of Transduction through Gradient Boosting and Application to Multi-scale Graph Neural Networks proceedings.neurips.cc/paper/2020/… github.com/delta2323/G… [NeurIPS 2020] Towards Deeper Graph Neural Networks with Differentiable Group Normalization arxiv.org/abs/2006.06… [ICML 2020 Workshop GRL+] A Note on Over-Smoothing for Graph Neural Networks arxiv.org/abs/2006.13… [ICML 2020] Bayesian Graph Neural Networks with Adaptive Connection Sampling arxiv.org/abs/2006.04… [ICML 2020] Continuous Graph Neural Networks arxiv.org/abs/1912.00… [ICML 2020] Simple and Deep Graph Convolutional Networks arxiv.org/abs/2007.02… github.com/chennnM/GCN… [KDD 2020] Towards Deeper Graph Neural Networks arxiv.org/abs/2007.09… github.com/mengliu1998… [ICLR 2020] Graph Neural Networks Exponentially Lose Expressive Power for Node Classification arxiv.org/abs/1905.10… github.com/delta2323/g…) [ICLR 2020] DropEdge: Towards Deep Graph Convolutional Networks on Node Classification openreview.net/forum?id=Hk… github.com/DropEdge/Dr… [ICLR 2020] PairNorm: Tackling Oversmoothing in GNNs openreview.net/forum?id=rk… github.com/LingxiaoSha…) [ICLR 2020] Measuring and Improving the Use of Graph Information in Graph Neural Networks openreview.net/forum?id=rk… github.com/yifan-h/CS-… [AAAI 2020] Measuring and Relieving the Over-smoothing Problem for Graph Neural Networks from the Topological View arxiv.org/abs/1909.03… Before 2020
[arXiv 2019] Revisiting Graph Neural Networks: All We Have is Low-Pass Filters arxiv.org/abs/1905.09… [NeurIPS 2019] Break the Ceiling: Stronger Multi-scale Deep Graph Convolutional Networks arxiv.org/abs/1906.02… [ICLR 2019] Predict then Propagate: Graph Neural Networks meet Personalized PageRank arxiv.org/abs/1810.05… github.com/klicperajo/… [ICCV 2019] DeepGCNs: Can GCNs Go as Deep as CNNs? arxiv.org/abs/1904.03…) github.com/lightaime/d… github.com/lightaime/d… [ICML 2018] Representation Learning on Graphs with Jumping Knowledge Networks arxiv.org/abs/1806.03… [AAAI 2018] Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learning arxiv.org/abs/1801.07…
