TCL: Transformer-based Dynamic Graph Modelling via Contrastive Learning
arXiv:2105.07944
Abstract
Dynamic graph modeling has recently attracted much attention due to its extensive applications in many real-world scenarios, such as recommendation systems, financial transactions, and social networks. Although many works have been proposed for dynamic graph modeling in recent years, effective and scalable models are yet to be developed. In this paper, we propose a novel graph neural network approach, called TCL, which deals with the dynamically-evolving graph in a continuous-time fashion and enables effective dynamic node representation learning that captures both the temporal and topology information. Technically, our model contains three novel aspects. First, we generalize the vanilla Transformer to temporal graph learning scenarios and design a graph-topology-aware transformer. Secondly, on top of the proposed graph transformer, we introduce a two-stream encoder that separately extracts representations from temporal neighborhoods associated with the two interaction nodes and then utilizes a co-attentional transformer to model inter-dependencies at a semantic level. Lastly, we are inspired by the recently developed contrastive learning and propose to optimize our model by maximizing mutual information (MI) between the predictive representations of two future interaction nodes. Benefiting from this, our dynamic representations can preserve high-level (or global) semantics about interactions and thus is robust to noisy interactions. To the best of our knowledge, this is the first attempt to apply contrastive learning to representation learning on dynamic graphs. We evaluate our model on four benchmark datasets for interaction prediction and experiment results demonstrate the superiority of our model.
References in corpus (12)
- A Simple Framework for Contrastive Learning of Visual Representations
- Inductive Representation Learning on Large Graphs
- Modeling polypharmacy side effects with graph convolutional networks
- DynGEM: Deep Embedding Method for Dynamic Graphs
- On Mutual Information Maximization for Representation Learning
- Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
- IntentGC: a Scalable Graph Convolution Framework Fusing Heterogeneous Information for Recommendation
- InfoGraph: Unsupervised and Semi-supervised Graph-Level Representation Learning via Mutual Information Maximization
- Temporal Graph Networks for Deep Learning on Dynamic Graphs
- SocialGCN: An Efficient Graph Convolutional Network based Model for Social Recommendation
- Deep Coevolutionary Network: Embedding User and Item Features for Recommendation
- DeepTrax: Embedding Graphs of Financial Transactions