![[2602.14049] UniST-Pred: A Robust Unified Framework for Spatio-Temporal Traffic Forecasting in Transportation Networks Under Disruptions](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.14049] UniST-Pred: A Robust Unified Framework for Spatio-Temporal Traffic Forecasting in Transportation Networks Under Disruptions
Summary
The article presents UniST-Pred, a novel framework for spatio-temporal traffic forecasting that effectively addresses disruptions in transportation networks, demonstrating strong predictive performance and interpretability.
Why It Matters
As urbanization increases, effective traffic management becomes crucial for reducing congestion and improving safety. UniST-Pred's ability to maintain performance under disruptions highlights its potential for real-world applications in intelligent transportation systems, making it a significant contribution to the field of machine learning and traffic forecasting.
Key Takeaways
- UniST-Pred decouples temporal modeling from spatial representation for improved forecasting.
- The framework is robust under severe network disruptions, ensuring reliability in real-world scenarios.
- It demonstrates competitive performance against established models while maintaining a lightweight design.
- The approach yields interpretable spatio-temporal representations, aiding in understanding traffic dynamics.
- Source code and datasets are publicly available, promoting further research and application.
Computer Science > Machine Learning arXiv:2602.14049 (cs) [Submitted on 15 Feb 2026] Title:UniST-Pred: A Robust Unified Framework for Spatio-Temporal Traffic Forecasting in Transportation Networks Under Disruptions Authors:Yue Wang, Areg Karapetyan, Djellel Difallah, Samer Madanat View a PDF of the paper titled UniST-Pred: A Robust Unified Framework for Spatio-Temporal Traffic Forecasting in Transportation Networks Under Disruptions, by Yue Wang and 3 other authors View PDF HTML (experimental) Abstract:Spatio-temporal traffic forecasting is a core component of intelligent transportation systems, supporting various downstream tasks such as signal control and network-level traffic management. In real-world deployments, forecasting models must operate under structural and observational uncertainties, conditions that are rarely considered in model design. Recent approaches achieve strong short-term predictive performance by tightly coupling spatial and temporal modeling, often at the cost of increased complexity and limited modularity. In contrast, efficient time-series models capture long-range temporal dependencies without relying on explicit network structure. We propose UniST-Pred, a unified spatio-temporal forecasting framework that first decouples temporal modeling from spatial representation learning, then integrates both through adaptive representation-level fusion. To assess robustness of the proposed approach, we construct a dataset based on an agent-based, microscop...
