CastFlow: Learning Role-Specialized Agentic Workflows for Time Series Forecasting

• URL: http://arxiv.org/abs/2604.27840v2
• Date: Mon, 04 May 2026 08:56:40 GMT
• Title: CastFlow: Learning Role-Specialized Agentic Workflows for Time Series Forecasting
• Abstract summary: We propose CastFlow, a dynamic agentic forecasting framework for large language models (LLMs)<n>CastFlow organizes the forecasting process into planning, action, forecasting, and reflection, establishing an agentic workflow.<n>Third, CastFlow adopts a role-specialized design that combines general-purpose reasoning with specialized numerical forecasting.
• Score: 74.29761960394848
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recently, large language models (LLMs) have shown great promise in time series forecasting. However, most existing LLM-based forecasting methods still follow a static generative paradigm that directly maps historical observations to future values in a single pass. Under this paradigm, forecasting is constrained by limited temporal pattern extraction, single-round acquisition of contextual features, one-shot forecast generation, and lack of support from ensemble forecasts. To address these limitations, in this work, we propose CastFlow, a dynamic agentic forecasting framework that enables multi-view temporal pattern extraction, multi-round contextual features acquisition, iterative forecast refinement, and forecasting with ensemble forecasts. First, CastFlow organizes the forecasting process into planning, action, forecasting, and reflection, establishing an agentic workflow. Second, this workflow is supported by a memory module that retrieves prior experience and a multi-view toolkit that constructs diagnostic evidence and provides a reliable ensemble forecast baseline. Third, CastFlow adopts a role-specialized design that combines general-purpose reasoning with specialized numerical forecasting. Under this design, a frozen LLM preserves general-purpose reasoning, while a fine-tuned domain-specific LLM performs evidence-guided numerical forecasting based on the ensemble forecast baseline, rather than from scratch. To optimize a fine-tuned domain-specific LLM, we further develop a two-stage workflow-oriented training that combines supervised fine-tuning (SFT) and reinforcement learning with verifiable rewards (RLVR). To evaluate the effectiveness of CastFlow, we conduct extensive experiments on diverse datasets and show that it achieves superior overall results against strong baselines. We hope that this work can serve as a step toward more adaptive and accurate time series forecasting.

Related papers

• CastFSR: A Fast--Slow--Reflect Agentic Reasoning Framework for Context-Aware Time Series Forecasting [55.05716348844572]
  CastFSR is an agentic framework that formulates context-aware forecasting as a Fast--Slow-Reflect-- workflow.<n>It retrieves contextual evidence, adaptively determines informative look-back windows, and reasons about how contexts reshape future dynamics.<n>It iteratively refines forecasts to ensure temporal, contextual, and domain consistency.
  arXiv  Detail & Related papers  (2026-08-04T02:21:36Z)
• Bridging the Last Mile of Time Series Forecasting with LLM Agents [0.6416429054645992]
  In real-world forecasting settings, a statistically plausible baseline is rarely the final forecast used in practice.<n>This practical stage remains underexplored in the forecasting literature.<n>We present an LLM-agent framework that sits on top of a forecasting backbone.
  arXiv  Detail & Related papers  (2026-06-01T17:03:40Z)
• LEAF: A Living Benchmark for Event-Augmented Forecasting [78.94019460534442]
  LEAF is the first living benchmark for event-augmented forecasting tasks.<n>We evaluate state-of-the-art proprietary and open-weight LLMs.<n>We find that LLMs achieve better performance on equities they confidently identify as more predictable.
  arXiv  Detail & Related papers  (2026-05-09T03:17:59Z)
• T-LLM: Teaching Large Language Models to Forecast Time Series via Temporal Distillation [7.6933817667680096]
  Time series forecasting plays a critical role in decision-making across many real-world applications.<n>We propose T-LLM, a temporal distillation framework that equips general-purpose language models with time series forecasting capability.
  arXiv  Detail & Related papers  (2026-02-02T10:40:27Z)
• Position: Beyond Model-Centric Prediction -- Agentic Time Series Forecasting [49.05788441962762]
  We argue for agentic time series forecasting (ATSF), which reframes forecasting as an agentic process composed of perception, planning, action, reflection, and memory.<n>We outline three representative implementation paradigms -- workflow-based design, agentic reinforcement learning, and a hybrid agentic workflow paradigm -- and discuss the opportunities and challenges that arise when shifting from model-centric prediction to agentic forecasting.
  arXiv  Detail & Related papers  (2026-02-02T08:01:11Z)
• Agentic Predictor: Performance Prediction for Agentic Workflows via Multi-View Encoding [56.565200973244146]
  Agentic Predictor is a lightweight predictor for efficient agentic workflow evaluation.<n>By learning to approximate task success rates, Agentic Predictor enables fast and accurate selection of optimal agentic workflow configurations.
  arXiv  Detail & Related papers  (2025-05-26T09:46:50Z)
• Generating ensembles of spatially-coherent in-situ forecasts using flow matching [0.3721016050171826]
  We propose a machine-learning-based methodology for in-situ weather forecast postprocessing.<n>Our Flow MAtching Postprocessing (FMAP) better represents the correlation structures of the observations distribution.<n>FMAP shows promising performance in experiments on the EUPPBench dataset.
  arXiv  Detail & Related papers  (2025-04-04T14:12:53Z)
• TimeXL: Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop [79.5773512667468]
  TimeXL is a multi-modal prediction framework that integrates a prototype-based time series encoder with three collaborating Large Language Models.<n>A reflection LLM compares the predicted values against the ground truth, identifying textual inconsistencies or noise.<n>This closed-loop workflow-prediction, critique (reflect), and refinement-continuously boosts the framework's performance and interpretability.
  arXiv  Detail & Related papers  (2025-03-02T20:40:53Z)
• FLODCAST: Flow and Depth Forecasting via Multimodal Recurrent Architectures [31.879514593973195]
  We propose a flow and depth forecasting model, trained to jointly forecast both modalities at once. We train the proposed model to also perform predictions for several timesteps in the future. We report benefits on the downstream task of segmentation forecasting, injecting our predictions in a flow-based mask-warping framework.
  arXiv  Detail & Related papers  (2023-10-31T16:30:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.