DynaSTy: A Framework for Spatio-Temporal Node Attribute Prediction in Dynamic Graphs

ICLR 2026 Conference Submission22386 Authors

Keywords: dynamic networks, node attribute prediction, graph neural networks, time series forecasting

TL;DR: This paper proposes a novel method for multistep, multidimensional node attribute prediction in dynamic graphs with evolving edges.

Abstract:

Accurate multi‐step forecasting of node‐level attributes on dynamic graphs is critical for applications ranging from financial trust networks to traffic monitoring. Existing spatio‐temporal graph neural networks typically assume a static adjacency, and seldom deal with multidimensional timeseries prediction. In this work, we propose an end‐to‐end dynamic edge‐biased spatio‐temporal model that ingests a multidimensional timeseries of node attributes and a timeseries of adjacency matrices, to predict multiple future steps of node attributes. At each time step, our transformer-based model injects the given adjacency as an adaptable attention bias, allowing the model to focus on relevant neighbors as the graph evolves. We further deploy a masked node/time pretraining objective that primes the encoder to reconstruct missing features, and train with scheduled sampling and a horizon‐weighted loss to mitigate compounding error over long horizons.
Unlike prior work, our model accommodates dynamic graphs that vary across input samples, enabling forecasting in multi-system settings such as brain networks across different subjects, financial systems in different contexts, or evolving social systems.
Empirical results demonstrate that our method outperforms strong STGCN, DCRNN, and MTGNN baselines by 10–20% on most datasets.

Primary Area: learning on graphs and other geometries & topologies

Submission Number: 22386