Graph Self-Supervised Learning: A Surveyieeexplore.ieee.org/abstract/document/9770382 PDF: https://arxiv.org/pdf/2103.00111.pdfarxiv.org/pdf/2103.00111.pdf 1. 背景与概括 图的深度学习作为一个热门领域引起了广泛的研究兴趣,但是由于当前研究大多集中在半监督或者监督学习上,存在标签依赖严重、泛...
E. Buchatskaya, C. Doersch, B. Avila Pires, Z. Guo, M. Ghesh- laghi Azar, B. Piot, k. kavukcuoglu, R. Munos, and M. Valko, “Bootstrap your own latent - a new approach to self-supervised learning,” in NeurIPS, vol. 33, 2020, pp. 21 271–21 284. ...
【图深度自监督学习Philips S. Yu团队重磅新作】Graph Self-Supervised Learning: A Survey,程序员大本营,技术文章内容聚合第一站。
Multi-stage self-supervised learning for graph convolutional networks on graphs with few labeled nodes. AAAI, 2020. Deep clustering for unsupervised learning of visual features. ECCV, 2018. When does self-supervision help graph convolutional networks?. ICML, 2020. Pre-training graph neural networks f...
图上的深度学习最近引起了人们的极大兴趣。然而,大多数工作都集中在(半)监督学习上,导致存在标签依赖重、泛化能力差和鲁棒性弱等缺点。为了解决这些问题,自监督学习 (SSL) 通过精心设计的借口任务提取信息知识,而不依赖于手动标签,已成为图数据的一种有前途和趋势的学习范式。与计算机视觉和自然语言处理等其他领域中...
https://github.com/LirongWu/awesome-graph-self-supervised-learning 近些年来,图上的深度学习在各种任务上取得了显著的成功,而这种成功在很大程度上依赖于海量的、精心标注的数据。然而,精确的标注通常非常昂贵和耗时。为了解决这个问题,自监督学习(Self-supervised Learning,SSL)正在成为一种全新的范式,通过精心设计的...
In this paper, we introduce a new self-supervised graph representation learning method DGB. DGB relies on two neural networks: online network and target network, and the input of each neural network is an augmentation of the initial graph. With the help of the bootstrapping process, the onlin...
Hierarchical Molecular Graph Self-Supervised Learning for property prediction Article Open access 17 February 2023 A knowledge-guided pre-training framework for improving molecular representation learning Article Open access 21 November 2023 Data availability All data used in this paper are publicly av...
Graph Self-Supervised Learning: A Survey-对比学习方法 AI知其然发表于图神经网络... A Comprehensive Survey on Graph Neural Networks简译 红豆绿豆芋...发表于浅究深度学... Robustness of deep learning models on graphs: A survey图深度学习鲁棒性综述 数据的小米...发表于图神经网络... Graph Attention...
2.2 GraphDA for Low-Resource Graph Learning 2.2.1 Graph Self-Supervised Learning 收auto-encoder 启发,图生成建模方法对输入的图进行数据扩充,然后通过从扩充的图中恢复特征/结构信息来学习模型。对于输入图,节点或边的特征被掩盖然后目标是通过 GNN 根据未屏蔽的信息重新恢复屏蔽的特征/结构[18][19][20][21...