AI Paper Atlas

Research map / Time series / Foundation models for time series

GTM: A General Time-series Model for Enhanced Representation Learning of Time-Series data

Cheng HE, Xu Huang, Gangwei Jiang, Zhaoyi Li, Defu Lian, Hong Xie, Enhong Chen, xijie liang, Zhengzengrong, Patrick Lee

ICLR 2026 · Time series › Foundation models for time series · approach: Time series representation learning · 1 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
pre-train GTM on UTSD-12G then fine-tune vs train from random initialization (ablation)GTM · long-term forecastingMSE: MSE reduction 0.5%-7.8%Improved
pre-train GTM vs random initialization (ablation)GTM · imputationMSE: MSE reduction 1.2%-11.7%Improved
pre-train GTM vs random initialization (ablation)GTM · anomaly detectionF1: +1.2% average F1Improved
add frequency-domain and time-granularity-aware modules (Fourier attention) (ablation)GTM · long-term forecastingMSEImproved
increase model depth and width in pre-training (ablation)GTM · ETTh1 forecastingMSEImproved
increase pre-training data size (ablation)GTM · ETTh1 and Weather forecastingMSEImproved
increase number of low-rank modules in Fourier Attention to 20 (ablation)GTM · inference latencylatency (s/item): stays below 0.1s/itemHurt
GTM zero-shot vs other TSFMsGTM · zero-shot forecastingaverage MSE: 0.392 -> 0.380 vs TIMER-1BMixed
GTM vs best baselines on classificationGTM · 10 classification datasetsaccuracyMixed
GTM vs baselines on imputationGTM · Electricity imputationMSE: 0.072 (PatchTST) vs 0.086Hurt
GTM vs baselines on anomaly detectionGTM · anomaly detection average F1F1: 86.72 -> 87.01Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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