Research map / Time series / Foundation models for time series
ConFlux: Multivariate Time Series in Flux, One Unified Forecast in Confluence
Shiyu Wang, Juntong Ni, Ziyi Zhang, Baichuan Mo, Xinyue Zhong, Chengxin Wang, Yuchen Fang, Zhou Ye, Yang Xiang
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 variate sorting (Fiedler-vector ordering) before variate patching (ablation) | ConFlux · ETTh1 zero-shot | MSE: 0.397 -> 0.378 | Improved |
| add variate sorting (ablation) | ConFlux · ETTm1 zero-shot | MSE: 0.325 -> 0.314 | Improved |
| add variate sorting (ablation) | ConFlux · Solar zero-shot | MSE: 0.165 -> 0.162 | Improved |
| train with variate patching (ablation) | ConFlux · Weather zero-shot | MSE: 0.270 -> 0.229 | Improved |
| train with variate patching (ablation) | ConFlux · ETTm1 zero-shot | MSE: 0.366 -> 0.314 | Improved |
| use variate patching at inference (ablation) | ConFlux · ETTh1 zero-shot | MSE: 0.390 -> 0.378 | Improved |
| use variate patching at inference (ablation) | ConFlux · Electricity zero-shot | MSE: 0.165 -> 0.165 | No clear effect |
| use ConFlux instead of Chronos-2 for zero-shot long-term forecasting | Chronos-2 · long-term datasets | MSE: average 12.5% reduction; up to 41.1% on Electricity | Improved |
| use ConFlux instead of Chronos-2 on short-term traffic | Chronos-2 · PEMS03/04/07/08 zero-shot | MSE: average 29.7% reduction; 39.3% on PEMS07 | Improved |
| in-domain fine-tuning vs multivariate Timer-XL | Timer-XL · long-term fine-tuned | MSE: average 30.7% reduction vs Timer-XL | Improved |
| in-domain fine-tuning ConFlux vs Chronos | Chronos · Electricity fine-tuned | MSE: 0.154 (Chronos) vs 0.159 (ConFlux) | Hurt |
| scale up model size (Mini to Large) with more data (ablation) | ConFlux · scalability | MAE | 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.
- Sundial — Sundial: A Family of Highly Capable Time Series Foundation Models (ICML 2025) · better on 2 of 2 shared results
- Timer-XL — Timer-XL: Long-Context Transformers for Unified Time Series Forecasting (ICLR 2025) · 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
- TiRex-2: Generalizing TiRex to Multivariate Data and Streaming (NeurIPS 2026 · similarity 0.90)
- OLinear: A Linear Model for Time Series Forecasting in Orthogonally Transformed Domain (NeurIPS 2025 · similarity 0.89)
- Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfer (ICML 2025 · similarity 0.88)
- TimeEmb: A Lightweight Static-Dynamic Disentanglement Framework for Time Series Forecasting (NeurIPS 2025 · similarity 0.88)
- xLSTM-Mixer: Multivariate Time Series Forecasting by Mixing via Scalar Memories (NeurIPS 2025 · similarity 0.88)
- Timer-XL: Long-Context Transformers for Unified Time Series Forecasting (ICLR 2025 · similarity 0.88)
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.
