Research map / Time series / Foundation models for time series
FlowState: Sampling-Rate‑Equivariant Time‑Series Forecasting
Lars Graf, Thomas Ortner, Stanisław Woźniak, Angeliki Pantazi
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 functional basis decoder instead of fixed linear decoder (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE: 0.754 -> 0.725 | Improved |
| add time noise to decoder sampling locations during training (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE: 0.740 -> 0.725 | Improved |
| use causal RevIN instead of standard RevIN (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE: 0.738 -> 0.725 | Improved |
| use all basis functions instead of first 128 at inference (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE: 0.726 -> 0.725 | No clear effect |
| use all basis functions instead of first 64 at inference (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE: 0.735 -> 0.725 | Improved |
| add sampling-rate adjustment mechanism (scale factor) (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE, CRPS | Improved |
| add parallel predictions during training (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE, CRPS | Improved |
| add output gate to SSM block (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE, CRPS | Improved |
| use complex SSM instead of real SSM (S4D-real init) (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE, CRPS | Improved |
| make Delta_E input-selective like S6 (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE, CRPS | Hurt |
| use Legendre basis instead of Fourier basis (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE, CRPS | Improved |
| forecast a quarter season (Legendre) instead of full season (Full-Legendre) (ablation) | FlowState-3M (2k) · GIFT-Eval (FlowState-3M, 2k context) | MASE, CRPS | Improved |
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.
- LightGTS-mini — LightGTS: A Lightweight General Time Series Forecasting Model (ICML 2025) · better on 9 of 16 shared results
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 2 of 2 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
- Multi-Scale Finetuning for Encoder-based Time Series Foundation Models (NeurIPS 2025 · similarity 0.87)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.86)
- Towards Neural Scaling Laws for Time Series Foundation Models (ICLR 2025 · similarity 0.86)
- Kairos: Toward Adaptive and Parameter-Efficient Time Series Foundation Models (NeurIPS 2026 · similarity 0.86)
- Learning Pattern-Specific Experts for Time Series Forecasting Under Patch-level Distribution Shift (NeurIPS 2025 · similarity 0.86)
- Flow Matching with Gaussian Process Priors for Probabilistic Time Series Forecasting (ICLR 2025 · 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.
