Combating Bilateral Edge Noise for Robust Link Prediction

Authors: Zhanke Zhou, Jiangchao Yao, Jiaxu Liu, Xiawei Guo, Quanming Yao, LI He, Liang Wang, Bo Zheng, Bo Han

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental we first conduct an empirical study to disclose that the edge noise bilaterally perturbs both input topology and target label, yielding severe performance degradation and representation collapse. ... Extensive experiments on six datasets and three GNNs with diverse noisy scenarios verify the effectiveness of our RGIB instantiations.
Researcher Affiliation Collaboration 1Hong Kong Baptist University 2CMIC, Shanghai Jiao Tong University 3Shanghai AI Laboratory 4 Alibaba Group 5 Tsinghua Unversity
Pseudocode Yes Algorithm 1 Hybrid graph augmentation.
Open Source Code Yes The code is publicly available at: https://github.com/tmlr-group/RGIB.
Open Datasets Yes Setup. 6 popular datasets and 3 types of GNNs are taken in the experiments.
Dataset Splits Yes The edge noise is generated based on Def. 3.1 after the commonly used data split where 85% edges are randomly selected for training, 5% as the validation set, and 10% for testing.
Hardware Specification Yes The software framework is the Pytorch [29], while the hardware is one NVIDIA RTX 3090 GPU.
Software Dependencies No The paper mentions 'Pytorch' as the software framework but does not specify a version number.
Experiment Setup Yes Setup. 6 popular datasets and 3 types of GNNs are taken in the experiments. The edge noise is generated based on Def. 3.1 after the commonly used data split where 85% edges are randomly selected for training, 5% as the validation set, and 10% for testing. The AUC is used as the evaluation metric as in [48, 55]. ... A further hyper-parameter study with grid search of the multiple λ in Eq. 3 of RGIB-SSL and Eq. 4 of RGIB-REP are illustrated in Fig. 7.