GraphNAS: Graph Neural Architecture Search with Reinforcement Learning

By Yang Gao, Hong Yang, Peng Zhang, Chuan Zhou, Yue Hu. 2019

121 citations recorded by the source corpus.

Graph Neural Networks (GNNs) have been popularly used for analyzing non-Euclidean data such as social network data and biological data. Despite their success, the design of graph neural networks requires a lot of manual work and domain knowledge. In this paper, we propose a Graph Neural Architecture Search method (GraphNAS for short) that enables automatic search of the best graph neural architecture based on reinforcement learning. Specifically, GraphNAS first uses a recurrent network to generate variable-length strings that describe the architectures of graph neural networks, and then trains the recurrent network with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation data set. Extensive experimental results on node classification tasks in both transductive and inductive learning settings demonstrate that GraphNAS can achieve consistently better performance on the Cora, Citeseer, Pubmed citation network, and protein-protein interaction network. On node classification tasks, GraphNAS can design a novel network architecture that rivals the best human-invented architecture in terms of test set accuracy.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Auto-GNN: Neural architecture search of graph neural networks (2019, methodology): “Following NAS framework in [27], we collect the set of attention functions as shown in Table 1 to compute coefficient a ( k ) ij .”
A Survey on Neural Architecture Search (2019, methodology): “ch for other types of networks. For instance, there are rst attempts to automatically optimize the architecture of autoencoders (Suganuma et al., 2018), transformers (So et al., 2019) and graph CNNs (Gao et al., 2019). In general, these techniques are increasingly being used to automate all componen”
Propagation Model Search for Graph Neural Networks (2020, methodology): “GraphNAS (Gao et al. 2019) and AGNN (Zhou et al. 2019) utilize reinforcement learning to search for GNN architectures.”
Graph Random Neural Networks for Semi-Supervised Learning on Graphs (2020, methodology): “• Eight graph convolutions: GCN [20], GAT [35], APPNP [21], Graph U-Net [12], SGC [39], MixHop [1], GMNN [31] and GrpahNAS [13].”
AutoGraph: Automated Graph Neural Network (2020, methodology): “ollows: min L val(w ();); (5) s.t. w () = argmin w L train(w;); (6) where w denotes the optimal parameters learned for the architecture in the training set. This is a bilevel optimization problem [6]. We propose an efﬁcient method to solve this problem based on the evolutionary algorithm. Each gener”
Probabilistic Dual Network Architecture Search on Graphs (2020, methodology): “3% lower on Citeseer compared to AGNN.”
Design Space for Graph Neural Networks (2020, methodology): “Architecture search techniques have been applied to GNNs [5, 51].”
Techniques for Automated Machine Learning (2019, methodology): “Recently, AutoDL has been developed to deal with segmentation [58; 91], graph-CNN [29; 100], GAN [33], and autoencoders [80].”
Evolutionary Architecture Search for Graph Neural Networks (2020, methodology): “• GraphNAS [15] first uses a recurrent network to generate variable-length strings that describe the architectures of graph neural networks, and then trains the recurrent network with reinforcement learning to maximize the embedding accuracy of the generated architectures.”