Todyformer: Towards Holistic Dynamic Graph Transformers with Structure-Aware Tokenization

TL;DR

Todyformer is a novel Transformer-based model for dynamic graphs that combines local and global encoding strategies to address over-squashing and over-smoothing, achieving superior performance on benchmark datasets.

Contribution

It introduces a structure-aware tokenization and an alternating encoding architecture to enhance dynamic graph modeling with Transformers.

Findings

Outperforms state-of-the-art methods on benchmark datasets

Effectively captures extensive temporal dependencies

Mitigates over-squashing and over-smoothing issues

Abstract

Temporal Graph Neural Networks have garnered substantial attention for their capacity to model evolving structural and temporal patterns while exhibiting impressive performance. However, it is known that these architectures are encumbered by issues that constrain their performance, such as over-squashing and over-smoothing. Meanwhile, Transformers have demonstrated exceptional computational capacity to effectively address challenges related to long-range dependencies. Consequently, we introduce Todyformer-a novel Transformer-based neural network tailored for dynamic graphs. It unifies the local encoding capacity of Message-Passing Neural Networks (MPNNs) with the global encoding of Transformers through i) a novel patchifying paradigm for dynamic graphs to improve over-squashing, ii) a structure-aware parametric tokenization strategy leveraging MPNNs, iii) a Transformer with temporal positional-encoding to capture long-range dependencies, and iv) an encoding architecture that alternates between local and global contextualization, mitigating over-smoothing in MPNNs. Experimental evaluations on public benchmark datasets demonstrate that Todyformer consistently outperforms the state-of-the-art methods for downstream tasks. Furthermore, we illustrate the underlying aspects of the proposed model in effectively capturing extensive temporal dependencies in dynamic graphs.