Do We Really Need Complicated Model Architectures For Temporal Networks?

Recurrent neural network (RNN) and self-attention mechanism (SAM) are the de facto methods to extract spatial-temporal information for temporal graph learning. Interestingly, we found that although both RNN and SAM could lead to a good performance, in practice neither of them is always necessary. In this paper, we propose \oure, a conceptually and technically simple architecture that consists of three components: \circled{1} a \emph{link-encoder} that is only based on multi-layer perceptrons (MLP) to summarize the information from temporal links, \circled{2} a \emph{node-encoder} that is only based on neighbor mean-pooling to summarize node information, and \circled{3} an MLP-based \emph{link classifier} that performs link prediction based on the outputs of the encoders. Despite its simplicity, \our attains an outstanding performance on temporal link prediction benchmarks with faster convergence and better generalization performance. These results motivate us to rethink the importance of simpler model architecture.