AI Paper Atlas

Research map / Time series / Causal and generative time series

Rethinking Time-Series Imputation as Conditional Inference along Temporal Evolution

Yu Fan, Yang Yang, Yufan Guo, Huazhong Yang, Pengjun Wang

ICML 2026 · Time series › Causal and generative time series · approach: Irregular series and imputation

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
use CBiT instead of the best baseline (ImputeFormer), 10% missingCBiT · ETTh1 10% missing (joint supervision)Masked MAE: 0.2369 -> 0.1658Improved
use CBiT instead of the best baseline (ImputeFormer), 10% missingCBiT · Power 10% missing (joint supervision)Masked MAE: 0.0485 -> 0.0336Improved
use CBiT instead of the best baseline (SAITS), 50% missingCBiT · WTH 50% missing (joint supervision)Masked MAE: 0.1514 -> 0.1487Improved
use CBiT instead of the best baseline (SSSD), 10% missingCBiT · ETTh1 10% missing (missing-only supervision)Masked MAE: 0.2821 -> 0.2002Improved
use CBiT instead of the best baseline (UNiTS), 10% missingCBiT · ETTh2 10% missing (missing-only supervision)Masked MAE: 0.1680 -> 0.1597Improved
use CBiT instead of ImputeFormer, 50% missingCBiT · Power 50% missing (missing-only supervision)Masked MAE: 0.0538 -> 0.0596No clear effect
add reconstruction loss on observed entries (joint vs missing-only supervision) (ablation)CBiT · ETTh1 10% missingMasked MAE: 0.2002 -> 0.1658Improved
add reconstruction loss on observed entries (joint vs missing-only supervision) (ablation)CBiT · Power 50% missingMasked MAE: 0.0596 -> 0.0557Improved
add reconstruction loss on observed entries (joint vs missing-only supervision) (ablation)CBiT · WTH 50% missingMasked MAE: 0.1777 -> 0.1487Improved
use CBiT instead of UNiTS in computeCBiT · Computational costFLOPs (G): 2.2620 -> 0.0586Improved
use CBiT instead of SAITS in parametersCBiT · Computational costParams (M): 0.2045 -> 0.2753Hurt

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