Research map / Time series / Long-term forecasting
ASTGI: Adaptive Spatio-Temporal Graph Interactions for Irregular Multivariate Time Series Forecasting
Xvyuan Liu, Xiangfei Qiu, Hanyin Cheng, Xingjian Wu, Chenjuan Guo, Bin Yang, Jilin Hu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use learnable time and channel embeddings instead of fixed non-parametric encodings (ablation) | ASTGI · MIMIC | MSE: 0.4057 -> 0.3909 | Improved |
| search neighbors in learned spatio-temporal coordinates instead of original timestamps (ablation) | ASTGI · PhysioNet | MSE: 0.3164 -> 0.3004 | Improved |
| add spatio-temporal displacement vector to interaction scoring and messages (ablation) | ASTGI · USHCN | MSE: 0.1930 -> 0.1607 | Improved |
| use interaction-weighted query fusion instead of mean pooling (ablation) | ASTGI · MIMIC | MSE: 0.8807 -> 0.3909 | Improved |
| use interaction-weighted query fusion instead of mean pooling (ablation) | ASTGI · Human Activity | MSE: 0.0870 -> 0.0412 | Improved |
| use learned coordinates, adaptive graph, relation-aware components (ablation) | ASTGI · Human Activity | MSE: 0.0421/0.0421/0.0418 -> 0.0412 | No clear effect |
| mean pooling in place of weighted fusion on USHCN (ablation) | ASTGI · USHCN | MSE: 0.1699 -> 0.1607 | Improved |
| increase number of propagation layers L too far (ablation) | ASTGI · sensitivity analysis | MSE | Hurt |
| increase number of candidate neighbors K (ablation) | ASTGI · sensitivity analysis | MSE | No clear effect |
| ASTGI vs second-best Hi-Patch across datasets | ASTGI · four IMTS datasets | MSE: approximately 6.04% MSE reduction | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GraFITi — GraFITi: Graphs for Forecasting Irregularly Sampled Time Series (AAAI 2024) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Irregular Multivariate Time Series Forecasting: A Transformable Patching Graph Neural Networks Approach (ICML 2024 · similarity 0.90)
- Graph-based Forecasting with Missing Data through Spatiotemporal Downsampling (ICML 2024 · similarity 0.87)
- Graph Neural Flows for Unveiling Systemic Interactions Among Irregularly Sampled Time Series (NeurIPS 2024 · similarity 0.87)
- HyperIMTS: Hypergraph Neural Network for Irregular Multivariate Time Series Forecasting (ICML 2025 · similarity 0.87)
- Structured Matrix Basis for Multivariate Time Series Forecasting with Interpretable Dynamics (NeurIPS 2024 · similarity 0.86)
- Towards Dynamic Spatial-Temporal Graph Learning: A Decoupled Perspective (AAAI 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
