DyG-Mamba: Continuous State Space Modeling on Dynamic Graphs
This work addresses the problem of uncovering evolutionary patterns in dynamic graphs for applications such as social recommendation and cancer detection, representing an incremental advance by adapting state space models to this domain.
The paper tackles dynamic graph modeling by translating it into a long-term sequence modeling problem, achieving state-of-the-art performance on most of 12 datasets for tasks like dynamic link prediction and node classification, with significant improvements in computational and memory efficiency.
Dynamic graph modeling aims to uncover evolutionary patterns in real-world systems, enabling accurate social recommendation and early detection of cancer cells. Inspired by the success of recent state space models in efficiently capturing long-term dependencies, we propose DyG-Mamba by translating dynamic graph modeling into a long-term sequence modeling problem. Specifically, inspired by Ebbinghaus' forgetting curve, we treat the irregular timespans between events as control signals, allowing DyG-Mamba to dynamically adjust the forgetting of historical information. This mechanism ensures effective usage of irregular timespans, thereby improving both model effectiveness and inductive capability. In addition, inspired by Ebbinghaus' review cycle, we redefine core parameters to ensure that DyG-Mamba selectively reviews historical information and filters out noisy inputs, further enhancing the model's robustness. Through exhaustive experiments on 12 datasets covering dynamic link prediction and node classification tasks, we show that DyG-Mamba achieves state-of-the-art performance on most datasets, while demonstrating significantly improved computational and memory efficiency. Code is available at https://github.com/Clearloveyuan/DyG-Mamba.