AI Paper Atlas

Research map / Time series / Foundation models for time series

Semantic-Enhanced Time-Series Forecasting via Large Language Models

Hao Liu, Zhang xiaoxing, Chun Yang, Xiaobin Zhu

ICLR 2026 · Time series › Foundation models for time series · approach: LLM-based time series forecasting · 6 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 cross-attention alignment with TSCC (ablation)SE-LLM baseline with BERT · Traffic (output 96)MSE: 0.472 -> 0.431Improved
add Time-Adapter on top of TSCC (ablation)SE-LLM with BERT · Traffic (output 96)MSE: 0.431 -> 0.412Improved
add Time-Adapter on top of TSCC (ablation)SE-LLM with Qwen2.5-0.5B · ECL (output 96)MSE: 0.166 -> 0.161Improved
add AM-VAE anomaly modeling (ablation)TSCC · ETTh1MSE: 0.393 -> 0.381Improved
add cross-domain attention instead of linear concatenation (ablation)TSCC · ETTh1MSE: 0.396 -> 0.381Improved
add gated fusion (ablation)TSCC · ETTh1MSE: 0.399 -> 0.381Improved
use LLM-derived semantic space instead of learnable matrix (ablation)TSCC · ETTh1MSE: 0.402 -> 0.381Improved
AM-VAE benefit depends on data regime (ablation)TSCC · various datasetsMSEMixed
use Time-Adapter instead of LoRA (ablation)Qwen2.5 with TSCC · ETTh1MSE: 0.396 -> 0.389Improved
add LoRA to TSCC baseline (ablation)Qwen2.5 with TSCC · ETTh1MSE: 0.381 -> 0.396Hurt
add Time-Adapter to TSCC baseline (ablation)Qwen2.5 with TSCC · ETTh1MSE: 0.381 -> 0.389Hurt
add Time-Adapter to TSCC baseline (ablation)Qwen2.5 with TSCC · WeatherMSE: 0.242 -> 0.229Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 5 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