Research map / Time series / Long-term forecasting
Local Geometry Attention for Time Series Forecasting under Realistic Corruptions
Dongbin Kim, Youngjoo Park, Woojin Jeong, Jaewook Lee
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 |
|---|---|---|---|
| replace scaled dot-product attention with LGA | PatchTST · ETTm1 combined severity 5 | MSE: 0.839 -> 0.734 | Improved |
| replace scaled dot-product attention with LGA | PatchTST · ETTm1 combined severity 3 | MSE: 0.614 -> 0.519 | Improved |
| replace scaled dot-product attention with LGA | PatchTST · Weather combined severity 5 | MSE: 0.491 -> 0.454 | Improved |
| replace scaled dot-product attention with LGA | PatchTST · Electricity combined severity 5 | MSE: 0.290 -> 0.284 | Improved |
| replace scaled dot-product attention with LGA | PatchTST · ETTm1 clean (severity 0) | MSE: 0.352 -> 0.351 | No clear effect |
| replace TimeMixer with PatchLGA | PatchLGA · Weather combined severity 5 | MSE: 0.576 -> 0.454 | Improved |
| replace standard attention with MoM robust attention | PatchTST · ETTm1 combined severity 5 | MSE: 0.839 -> 1.016 | Hurt |
| replace standard attention with Elliptical attention | PatchTST · ETTm1 combined severity 5 | MSE: 0.839 -> 0.880 | Hurt |
| replace SDP with LGA in CATS (cross-attention) | CATS · ETTm1 combined severity 5 | MSE: 1.102 -> 1.037 | Improved |
| replace SDP with LGA in iTransformer (channel attention) | iTransformer · ETTm1 combined severity 5 | MSE: 1.309 -> 1.266 | Improved |
| replace SDP with LGA under level shift only | PatchTST · ETTm1 level shift severity 5 | MSE: 0.848 -> 0.718 | Improved |
| replace SDP with LGA under spike corruption only | PatchTST · ETTm1 spike severity 5 | MSE: 0.374 -> 0.372 | No clear effect |
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.
- TimeMixer — TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (ICLR 2024) · better on 7 of 7 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
- SynTSBench: Rethinking Temporal Pattern Learning in Deep Learning Models for Time Series (NeurIPS 2025 · similarity 0.85)
- TimeBase: The Power of Minimalism in Efficient Long-term Time Series Forecasting (ICML 2025 · similarity 0.85)
- A Closer Look at Transformers for Time Series Forecasting: Understanding Why They Work and Where They Struggle (ICML 2025 · similarity 0.85)
- How Different from the Past? Spatio-Temporal Time Series Forecasting with Self-Supervised Deviation Learning (NeurIPS 2025 · similarity 0.85)
- Introducing Spectral Attention for Long-Range Dependency in Time Series Forecasting (NeurIPS 2024 · similarity 0.85)
- Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenization (ICML 2025 · similarity 0.85)
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.
