TimeXL: Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915879009124352 |
|---|---|
| author | Jiang, Yushan Yu, Wenchao Lee, Geon Song, Dongjin Shin, Kijung Cheng, Wei Liu, Yanchi Chen, Haifeng |
| author_facet | Jiang, Yushan Yu, Wenchao Lee, Geon Song, Dongjin Shin, Kijung Cheng, Wei Liu, Yanchi Chen, Haifeng |
| contents | Time series analysis provides essential insights for real-world system dynamics and informs downstream decision-making, yet most existing methods often overlook the rich contextual signals present in auxiliary modalities. To bridge this gap, we introduce TimeXL, a multi-modal prediction framework that integrates a prototype-based time series encoder with three collaborating Large Language Models (LLMs) to deliver more accurate predictions and interpretable explanations. First, a multi-modal prototype-based encoder processes both time series and textual inputs to generate preliminary forecasts alongside case-based rationales. These outputs then feed into a prediction LLM, which refines the forecasts by reasoning over the encoder's predictions and explanations. Next, a reflection LLM compares the predicted values against the ground truth, identifying textual inconsistencies or noise. Guided by this feedback, a refinement LLM iteratively enhances text quality and triggers encoder retraining. This closed-loop workflow-prediction, critique (reflect), and refinement-continuously boosts the framework's performance and interpretability. Empirical evaluations on four real-world datasets demonstrate that TimeXL achieves up to 8.9% improvement in AUC and produces human-centric, multi-modal explanations, highlighting the power of LLM-driven reasoning for time series prediction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_01013 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | TimeXL: Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop Jiang, Yushan Yu, Wenchao Lee, Geon Song, Dongjin Shin, Kijung Cheng, Wei Liu, Yanchi Chen, Haifeng Machine Learning Time series analysis provides essential insights for real-world system dynamics and informs downstream decision-making, yet most existing methods often overlook the rich contextual signals present in auxiliary modalities. To bridge this gap, we introduce TimeXL, a multi-modal prediction framework that integrates a prototype-based time series encoder with three collaborating Large Language Models (LLMs) to deliver more accurate predictions and interpretable explanations. First, a multi-modal prototype-based encoder processes both time series and textual inputs to generate preliminary forecasts alongside case-based rationales. These outputs then feed into a prediction LLM, which refines the forecasts by reasoning over the encoder's predictions and explanations. Next, a reflection LLM compares the predicted values against the ground truth, identifying textual inconsistencies or noise. Guided by this feedback, a refinement LLM iteratively enhances text quality and triggers encoder retraining. This closed-loop workflow-prediction, critique (reflect), and refinement-continuously boosts the framework's performance and interpretability. Empirical evaluations on four real-world datasets demonstrate that TimeXL achieves up to 8.9% improvement in AUC and produces human-centric, multi-modal explanations, highlighting the power of LLM-driven reasoning for time series prediction. |
| title | TimeXL: Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2503.01013 |