Research map / Time series / Long-term forecasting
Learning Recursive Multi-Scale Representations for Irregular Multivariate Time Series Forecasting
Boyuan Li, Zhen Liu, Yicheng Luo, Qianli Ma
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 |
|---|---|---|---|
| add ReIMTS recursive multi-scale wrapper to GraFITi | GraFITi · MIMIC-IV | MSE x10^-1: 2.39 -> 1.79 | Improved |
| add ReIMTS to GraFITi | GraFITi · MIMIC-III | MSE x10^-1: 4.08 -> 4.07 | No clear effect |
| add ReIMTS to PrimeNet | PrimeNet · Human Activity | MSE x10^-1: 26.84 -> 0.82 | Improved |
| add ReIMTS to six IMTS backbones | six IMTS backbones · five IMTS datasets | MSE: average 27.1% error reduction | Improved |
| add ReIMTS to GRU-D | GRU-D · PhysioNet'12 | MSE x10^-1: 3.25 -> 3.25 | No clear effect |
| split subsamples by time period instead of observation count (ablation) | ReIMTS (GraFITi) · MIMIC-IV | MSE x10^-1: 2.36 -> 1.79 | Improved |
| add recursive splitting into subsamples (vs original sample only) (ablation) | ReIMTS (GraFITi) · MIMIC-III | MSE x10^-1: 4.99 -> 4.07 | Improved |
| use irregularity-aware fusion instead of simple addition (ablation) | ReIMTS (GraFITi) · USHCN | MSE x10^-1: 1.89 -> 1.66 | Improved |
| use irregularity-aware fusion instead of simple addition (ablation) | ReIMTS (GraFITi) · PhysioNet'12 | MSE x10^-1: 2.79 -> 2.83 | Hurt |
| add fusion of global and local representations (vs none) (ablation) | ReIMTS (GraFITi) · MIMIC-III | MSE x10^-1: 4.77 -> 4.07 | Improved |
| increase scale levels from 2 to 3 (ablation) | ReIMTS (GraFITi) · Human Activity | MSE x10^-1: 0.42 -> 0.71 | Hurt |
| increase scale levels from 2 to 3 (ablation) | ReIMTS (GraFITi) · MIMIC-III | MSE x10^-1: 4.09 -> 4.07 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- GraFITi — GraFITi: Graphs for Forecasting Irregularly Sampled Time Series (AAAI 2024) · better on 5 of 5 shared results
- PrimeNet — PrimeNet: Pre-training for Irregular Multivariate Time Series (AAAI 2023) · better on 1 of 1 shared result
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
- HyperIMTS: Hypergraph Neural Network for Irregular Multivariate Time Series Forecasting (ICML 2025 · similarity 0.90)
- TimeCHEAT: A Channel Harmony Strategy for Irregularly Sampled Multivariate Time Series Analysis (AAAI 2025 · similarity 0.89)
- IMTS is Worth Time $\times$ Channel Patches: Visual Masked Autoencoders for Irregular Multivariate Time Series Prediction (ICML 2025 · similarity 0.89)
- Time-IMM: A Dataset and Benchmark for Irregular Multimodal Multivariate Time Series (NeurIPS 2025 · similarity 0.88)
- Irregular Multivariate Time Series Forecasting: A Transformable Patching Graph Neural Networks Approach (ICML 2024 · similarity 0.87)
- HDT: Hierarchical Discrete Transformer for Multivariate Time Series Forecasting (AAAI 2025 · similarity 0.87)
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.
