AI Paper Atlas

Research map / Time series / Foundation models for time series

Words Towards Explainability: Caption Label-Free Learning via Dual Loop Agentic Time Series Captioning

Difei Hou, Jiaqi Yue, Chunhui Zhao

ICML 2026 · Time series › Foundation models for time series · approach: Time series classification

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 RL optimization with downstream proxy rewardQwen2.5-VL-7B as Agent · Three-phase flow fault diagnosisdiagnosis accuracy: 0.3976 -> 0.9743Improved
add RL optimization with downstream proxy rewardQwen2.5-7B as Agent · Three-phase flow fault diagnosisdiagnosis accuracy: 0.3647 -> 0.8744Improved
add RL optimization with proxy rewardQwen2.5-VL-7B as Agent · Caption quality FActScoreFActScore: 0.081 -> 0.913Improved
add toolset and reflective agent workflow (no fine-tuning) (ablation)GPT-5.2 · Caption quality ROUGE-LROUGE-L: 0.247 -> 0.318Improved
add toolset and reflective agent workflow (no fine-tuning) (ablation)GPT-5.2 · Caption quality FActScoreFActScore: 0.738 -> 0.813Improved
add toolset and reflective agent workflow (no fine-tuning) (ablation)GPT-4o · Caption quality FActScoreFActScore: 0.711 -> 0.820Improved
add toolset and agent workflow without fine-tuning (ablation)Qwen2.5-7B · Caption quality FActScoreFActScore: 0.102 vs 0.450-0.738 for LLM baselinesNo clear effect
train on forecasting proxy instead of classification proxyOurs · Forecasting MAEMAE: 0.42 -> 0.39Improved
use captions from diagnosis-trained agent for forecasting instead of GPT-4o captionsGPT-4o captions · Forecasting MAEMAE: 0.61 -> 0.42Improved
apply method to gas turbine dataQwen2.5-VL-7B · Gas turbine fault diagnosisdiagnosis accuracy: 0.2314 -> 0.8872Improved
apply method to medical retinal signal dataQwen2.5-VL-7B · Disease diagnosisaccuracy: 0.4213 -> 0.8977Improved
apply method to general time-series reasoning dataQwen2.5-VL-7B · Classificationaccuracy: 0.4601 -> 0.8343Improved

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