Introduction Generative self-supervised learning (SSL),尤其是masked autoencoders在图数据挖掘中展现出巨大潜力。本文研究了异构图上的生成式SSL问题,并提出了一种异构图掩码自编码器模型——HGMAE。针对异构图,提出以下训练策略,包括基于元路径的边缘重建,节点属性恢复,以及对节点位置信息进行编码以进行位置特征预测。
自动编码器可以被认为是Kramer提出的主成分分析的非线性泛化。它的主要作用是降维或者说是特征学习,这种方式允许我们使用数据本身来进行特征提取和训练,而不依赖于人工标签。 在面对许多异构图任务时,尤其是生…
Recently, the masked autoencoder (MAE) has been shown to effectively pre-train Vision Transformers (ViT) to improve their representational capabilities. Methods: In this paper, we investigated a self-pre-training paradigm for serial SEM images with MAE to implement downstre...
Masked autoencoders are scalable vision learners. In CVPR, pages 16000–16009, June 2022. [23] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In CVPR, pages 770–778, 2016. [24] Geoffrey Hinton, Oriol Vinyals...
Specifically, we first propose a heterogeneous graph generative learning enhanced contrastive paradigm. This paradigm includes: 1) A contrastive view augmentation strategy by using a masked autoencoder. 2) Position-aware and semantics-aware positive sample sampling strategy for generating hard negative ...
Heterogeneous graph attention network for drug-target interaction prediction, 2022; pp. 1166–1176. https://doi.org/10.1145/3511808.3557346 Yang B, Chen H. Predicting circRNA-drug sensitivity associations by learning multimodal networks using graph auto-encoders and attention mechanism. Briefings Bioinf....
The first part encodes network-specific drug and target features from heterogeneous drug and gene/protein networks through sequential graph attention networks (GAT). The combined drug and target features (Fdt) are fed into the second part to learn the updated Fdt by the five transformer encoders....
Heterogeneous graph attention network for drug-target interaction prediction, 2022; pp. 1166–1176. https://doi.org/10.1145/3511808.3557346 Yang B, Chen H. Predicting circRNA-drug sensitivity associations by learning multimodal networks using graph auto-encoders and attention mechanism. Briefings Bioinf....
The meta-paths are consistent between in meta-path view based graph encoder and in contrastive learning. However, the connections among different meta-paths may vary greatly. The two nodes may have a large number of connections on just one type of meta-paths. We normalize the neighbor matrix ...
d HeteroSGT 将 RWR 序列作为输入,并生成子图表示 h_{S G 3.1 Preliminaries 3.2 News Heterogeneous Graph Construction 新闻 异构图构造 3.3 Dual-attention News Embedding Module 双注意力新闻嵌入模块 3.4 RWR-based Heterogeneous Subgraph Sampling 基于RWR的异构子图采样 c 构建异构图H_G对新闻、实体和主题之间...