paper-with-me

홈 › Papers

FDN: Interpretable Spatiotemporal Forecasting with Future Decomposition Networks

2026-06-23 · Nicholas Majeske, Ariful Azad arxiv

Spatiotemporal systems comprise a collection of spatially distributed yet interdependent entities each generating unique dynamic signals. Highly sophisticated methods have been proposed in recent years delivering state-of-the-art (SOTA) forecasts but few have focused on interpretability. To address this, we propose the Future Decomposition Network (FDN), a novel forecast model capable of (a) providing interpretable predictions through classification (b) revealing latent activity patterns in the target time-series and (c) delivering forecasts competitive with SOTA methods at a fraction of their memory and runtime cost. We conduct comprehensive analyses on FDN for multiple datasets from hydrologic, traffic, and energy systems, demonstrating its improved accuracy and interpretability.

📄 PDF Abstract BibTeX arXiv:2606.25201

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Interpretable Air Pollution Forecasting by Physics-Guided Spatiotemporal Decoupling

2025-11-25 · Zhiguo Zhang, Xiaoliang Ma, Daniel Schlesinger arxiv

Accurate and interpretable air pollution forecasting is crucial for public health, but most models face a trade-off between performance and interpretability. This study proposes a physics-guided, interpretable-by-design …

GeoPro-Net: Learning Interpretable Spatiotemporal Prediction Models through Statistically-Guided Geo-Prototyping

2024-12-19 · Bang An, Xun Zhou, Zirui Zhou, Ronilo Ragodos 외

The problem of forecasting spatiotemporal events such as crimes and accidents is crucial to public safety and city management. Besides accuracy, interpretability is also a key requirement for spatiotemporal forecasting m…

FUTURE-VLA: Forecasting Unified Trajectories Under Real-time Execution

2026-02-05 · Jingjing Fan, Yushan Liu, Shoujie Li, Botao Ren 외 arxiv

General vision-language models increasingly support unified spatiotemporal reasoning over long video streams, yet deploying such capabilities on robots remains constrained by the prohibitive latency of processing long-ho…

Graph-based Forecasting with Missing Data through Spatiotemporal Downsampling

2024-02-16 · Ivan Marisca, Cesare Alippi, Filippo Maria Bianchi

Given a set of synchronous time series, each associated with a sensor-point in space and characterized by inter-series relationships, the problem of spatiotemporal forecasting consists of predicting future observations f…

Missing ValuesTime Series

Unifying Physics- and Data-Driven Modeling via Novel Causal Spatiotemporal Graph Neural Network for Interpretable Epidemic Forecasting

2025-04-07 · Shuai Han, Lukas Stelz, Thomas R. Sokolowski, Kai Zhou 외

Accurate epidemic forecasting is crucial for effective disease control and prevention. Traditional compartmental models often struggle to estimate temporally and spatially varying epidemiological parameters, while deep l…

Graph Neural Network