AI Paper Atlas

Research map / Time series / Long-term forecasting

Tackling Time-Series Forecasting Generalization via Mitigating Concept Drift

Zhiyuan Zhao, Haoxin Liu, B. Aditya Prakash

ICLR 2026 · Time series › Long-term forecasting · approach: Attention for time series forecasting

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 changedWhereMeasuredEffect
add ShifTS (SAM plus RevIN) to forecasting backboneCrossformer · ILI (avg over horizons)MSE improvement: 81.9% MSE improvementImproved
add ShifTS to forecasting backboneiTransformer · ILI (avg over horizons)MSE improvement: 13.8% MSE improvementImproved
add ShifTS to forecasting backboneiTransformer · ETTm1 (avg over horizons)MSE improvement: 5.1% MSE improvementImproved
add ShifTS to older backbonesInformer · ETTh1 (avg over horizons)MSE improvement: 90.7% MSE improvementImproved
add ShifTS to TimeMixerTimeMixer · ETTm1 (avg over horizons)MSE improvement: 2.3% MSE improvementImproved
use ShifTS instead of concept drift baselines (GroupDRO, IRM, VREx, EIIL)Crossformer · ILIMSE: 2.036 (best concept drift baseline, EIIL) -> 0.668Improved
use invariant-learning concept drift methods alone without temporal shift handlingCrossformer · ExchangeMSE: 0.819 -> 0.821-0.846No clear effect
replace RevIN with SAN inside ShifTS (ablation)ShifTS on Crossformer · Exchange (avg)MSE: 0.470 -> 0.407Improved
mitigate both temporal shift and concept drift versus only one (ablation)ShifTS · Exchange horizon 96MSEImproved
add concept drift mitigation (SAM) to models with built-in norm/denorm (ablation)PatchTST, iTransformer · Exchange horizon 96MSEImproved
add temporal shift mitigation (RevIN) to a model without normalization (ablation)Crossformer · Exchange horizon 96MSEImproved
higher mutual information between horizon exogenous features and targetShifTS · six datasetsperformance gain: p=0.012Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy