Timezone: »
Self-supervised learning (SSL) of graph neural networks is emerging as a promising way of leveraging unlabeled data. Currently, most methods are based on contrastive learning adapted from the image domain, which requires view generation and a sufficient number of negative samples. In contrast, existing predictive models do not require negative sampling, but lack theoretical guidance on the design of pretext training tasks. In this work, we propose the LaGraph, a theoretically grounded predictive SSL framework based on latent graph prediction. Learning objectives of LaGraph are derived as self-supervised upper bounds to objectives for predicting unobserved latent graphs. In addition to its improved performance, LaGraph provides explanations for recent successes of predictive models that include invariance-based objectives. We provide theoretical analysis comparing LaGraph to related methods in different domains. Our experimental results demonstrate the superiority of LaGraph in performance and the robustness to decreasing of training sample size on both graph-level and node-level tasks.
Author Information
Yaochen Xie (Texas A&M University)
Zhao Xu (Texas A&M University)
Shuiwang Ji (Texas A&M University)
Related Events (a corresponding poster, oral, or spotlight)
-
2022 Poster: Self-Supervised Representation Learning via Latent Graph Prediction »
Tue. Jul 19th through Wed the 20th Room Hall E #207
More from the Same Authors
-
2023 : A new perspective on building efficient and expressive 3D equivariant graph neural networks »
weitao du · Yuanqi Du · Limei Wang · Dieqiao Feng · Guifeng Wang · Shuiwang Ji · Carla Gomes · Zhiming Ma -
2023 Poster: Group Equivariant Fourier Neural Operators for Partial Differential Equations »
Jacob Helwig · Xuan Zhang · Cong Fu · Jerry Kurtin · Stephan Wojtowytsch · Shuiwang Ji -
2023 Poster: Graph Mixup with Soft Alignments »
Hongyi Ling · Zhimeng Jiang · Meng Liu · Shuiwang Ji · Na Zou -
2023 Poster: Efficient Approximations of Complete Interatomic Potentials for Crystal Property Prediction »
Yuchao Lin · Keqiang Yan · Youzhi Luo · Yi Liu · Xiaoning Qian · Shuiwang Ji -
2023 Poster: Efficient and Equivariant Graph Networks for Predicting Quantum Hamiltonian »
Haiyang Yu · Zhao Xu · Xiaofeng Qian · Xiaoning Qian · Shuiwang Ji -
2022 Poster: Generating 3D Molecules for Target Protein Binding »
Meng Liu · Youzhi Luo · Kanji Uchino · Koji Maruhashi · Shuiwang Ji -
2022 Poster: GraphFM: Improving Large-Scale GNN Training via Feature Momentum »
Haiyang Yu · Limei Wang · Bokun Wang · Meng Liu · Tianbao Yang · Shuiwang Ji -
2022 Spotlight: GraphFM: Improving Large-Scale GNN Training via Feature Momentum »
Haiyang Yu · Limei Wang · Bokun Wang · Meng Liu · Tianbao Yang · Shuiwang Ji -
2022 Oral: Generating 3D Molecules for Target Protein Binding »
Meng Liu · Youzhi Luo · Kanji Uchino · Koji Maruhashi · Shuiwang Ji -
2021 Poster: On Explainability of Graph Neural Networks via Subgraph Explorations »
Hao Yuan · Haiyang Yu · Jie Wang · Kang Li · Shuiwang Ji -
2021 Spotlight: On Explainability of Graph Neural Networks via Subgraph Explorations »
Hao Yuan · Haiyang Yu · Jie Wang · Kang Li · Shuiwang Ji -
2021 Poster: GraphDF: A Discrete Flow Model for Molecular Graph Generation »
Youzhi Luo · Keqiang Yan · Shuiwang Ji -
2021 Spotlight: GraphDF: A Discrete Flow Model for Molecular Graph Generation »
Youzhi Luo · Keqiang Yan · Shuiwang Ji