Riemannian Liquid Spatio-Temporal Graph Network
Open Access
- 12 April 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 463-474
Abstract
Liquid Time-Constant networks (LTCs), a type of continuous-time graph neural network, excel at modeling irregularly-sampled dynamics but are fundamentally confined to Euclidean space. This limitation introduces significant geometric distortion when representing real-world graphs with inherent non-Euclidean structures (e.g., hierarchies and cycles), degrading representation quality. To overcome this limitation, we introduce the Riemannian Liquid Spatio-Temporal Graph Network (RLSTG), a framework that unifies continuous-time liquid dynamics with the geometric inductive biases of Riemannian manifolds. RLSTG models graph evolution through an Ordinary Differential Equation (ODE) formulated directly on a curved manifold, enabling it to faithfully capture the intrinsic geometry of both structurally static and dynamic spatio-temporal graphs. Moreover, we provide rigorous theoretical guarantees for RLSTG, extending stability theorems of LTCs to the Riemannian domain and quantifying its expressive power via state trajectory analysis. Extensive experiments on real-world benchmarks demonstrate that, by combining advanced temporal dynamics with a Riemannian spatial representation, RLSTG achieves superior performance on graphs with complex structures. Project Page: https://rlstg.github.ioKeywords
All Related Versions
This publication has 13 references indexed in Scilit:
- HGWaveNet: A Hyperbolic Graph Neural Network for Temporal Link Prediction2023
- Closed-form continuous-time neural networksNature Machine Intelligence, 2022
- Hyperbolic Spatial Temporal Graph Convolutional Networks2022
- Discrete-time Temporal Network Embedding via Implicit Hierarchical Learning in Hyperbolic Space2021
- Spatial-Temporal Graph ODE Networks for Traffic Flow Forecasting2021
- Predicting Dynamic Embedding Trajectory in Temporal Interaction Networks2019
- Attention Based Spatial-Temporal Graph Convolutional Networks for Traffic Flow ForecastingProceedings of the AAAI Conference on Artificial Intelligence, 2019
- Simplicial closure and higher-order link predictionProceedings of the National Academy of Sciences, 2018
- Implicit-Explicit Variational Integration of Highly Oscillatory ProblemsMultiscale Modeling & Simulation, 2009
- A Course in Metric Geometry2001