AI Paper Atlas

Research map / Time series / Foundation models for time series

TimeOmni-VL: Unified Models for Time Series Understanding and Generation

Tong Guan, Sheng Pan, Johan Barthelemy, Zhao Li, Yujun Cai, Cesare Alippi, Ming Jin, Shirui Pan

ICML 2026 · Time series › Foundation models for time series · approach: LLM-based time series forecasting · 0 citations (OpenAlex)

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
replace the heatmap TS-image with the TS2I periodic-cycle layout (ablation)TimeOmni-VL · GIFT-Eval forecasting and imputationnMASE: heatmap variant yields nMASE worse than the Naive baseline on nearly all tasksImproved
add generation CoT from the understanding model as context (ablation)TimeOmni-VL · GIFT-Eval forecasting and imputationnMASE: average 8.2% increase in nMASE without CoTImproved
post-train Bagel-7B on TSUMM-Suite for forecastingBagel-7B · GIFT-Eval forecasting, short-termnMASE: 16.303 -> 0.878Improved
post-train Bagel-7B on TSUMM-Suite for imputationBagel-7B · GIFT-Eval imputation, mask [0.1,0.2)nMASE: 17.411 -> 0.713Improved
post-train Bagel on TS-image understanding QA3Bagel-7B · TSUMM-Suite understanding QA3score: 0.012 -> 0.931Improved
post-train Bagel on TS-image understanding QA5Bagel-7B · TSUMM-Suite understanding QA5score: 0 -> 0.667Improved
compare TimeOmni-VL with VisionTS++ retrained on the same 5k samples, forecasting averageTimeOmni-VL · GIFT-Eval forecasting averagenMASE: 0.900 -> 0.826Improved
compare TimeOmni-VL with VisionTS++ retrained on the same 5k samples, imputation averageTimeOmni-VL · GIFT-Eval imputation averagenMASE: 0.827 -> 0.810Improved
compare TimeOmni-VL with VisionTS++ (original), long-term forecastingTimeOmni-VL · GIFT-Eval forecasting, long-termnMASE: 0.690 -> 0.784Hurt
compare TimeOmni-VL with TimeOmni-1 on reasoning Task1TimeOmni-VL · TSR-Suite Task1accuracy: 87.7 -> 84.0Hurt
compare TimeOmni-VL with linear interpolation, imputation mask [0.4,0.5]TimeOmni-VL · GIFT-Eval imputation, mask [0.4,0.5]nMASE: 0.968 -> 0.927No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 4 on most 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

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