GCN等工作能很好的学节点emb,早起GCN在图的变换域进行卷积操作,如graph Fourier transformation(图傅里叶变换)。为了降低计算复杂度,有研究学者提出polynomial spectral filters(多项式变换域过滤器),之后进一步简化为线性过滤器。与此同时,有工作直接在graph 空间域上进行卷积。之后,在卷积操作时,引入注意力机制,自适应...
However, learning representations that disentangle the latent factors poses great challenges and remains largely unexplored in the literature of graph neural networks. In this paper, we introduce the disentangled graph convolutional network (DisenGCN) to learn disentangled node representations. In particular...
Code:https://github.com/xiangwang1223/disentangled_graph_collaborative_filtering 这篇文章主要借鉴了ICML 2019的一篇文章的工作:Disentangled Graph Convolutional Networks。 动机 这篇文章主要是将ICML 2019的工作应用在了推荐上,用户-物品交互关系的建模发展过程可以概括为① 单个ID(用户、物品)的embedding ② 融入个...
Intent-aware Recommendation via Disentangled Graph Contrastive Learning论文阅读笔记 Abstract 存在的问题: 如何学习复杂和多样的意图,特别是当用户的行为在现实中通常不充分时 是不同的行为具有不同的意图分布,因此如何建立它们之间的关系,以建立一个更可解释的推荐系统。 本文方法: 在本文中,我们提出了通过解耦...
Official Code for "Disentangled Multi-Relational Graph Convolutional Network for Pedestrian Trajectory Prediction (AAAI 2021)" ihbae.com/publication/dmrgcn/ Topics deep-learning multi-agent autonomous-vehicles human-trajectory-prediction motion-forecasting disentangled-graph-neural-networks aaai2021 Resource...
More recently, graph convolutional network (GCN), a powerful deep representation learning method for graph-structured data, has achieved considerable attention in the field of FC analysis (Ktena et al., 2018, Yao et al., 2021). Most of the existing GCN methods learn the node (i.e., ROI...
Recommender systems aim to dig out the potential interests of users and find out items that might be connected with target users. Accuracy of the recommendation list is crucial for user-oriented applications. Many knowledge-based approaches combine graph
Many studies have shown that generative adversarial networks (GANs) can discover semantics at various levels of abstraction, yet GANs do not provide an intuitive way to show how they understand and control semantics. In order to identify interpretable di
Recently, Graph Convolutional Networks (GCNs) have shown remarkable success in the node classification task, due to the ability to aggregate neighborhood information and propagate supervised signals over the graph. However, most GCN-style models require relatively sufficient labeled data, which are not ...
Disentangled Graph Convolutional NetworksJianxin MaPeng CuiKun KuangXin WangWenwu ZhuPMLRInternational Conference on Machine Learning