cs.AINov 13, 2026

FlexCast: Flexible Context Learning for Time Series Foundation Models

Authors: Zuyang Shen, Luqi Gong, Fanda Fan, Bingsong Wang, Xiang Ao, Yahui Wei, Minghao Liu, Chunjie Luo, 

Organizations: Institute of Computing Technology, CAS · Zhejiang Lab · Beijing University of Posts and Telecommunications · Beijing Jiaotong University · 

Abstract

Time series foundation models (TSFMs) based on autoregressive (AR) next-token prediction entangle two factors: variable-context conditioning and prediction-unit factorization with autoregressive rollout. Strong AR performance therefore does not reveal whether generalization stems from the flexible training distribution, the local prediction target, the left-to-right inference process, or their combination. We propose FlexCast, a flexible context-target learning framework that disentangles these factors. FlexCast reformulates forecasting as variable-length masked prediction over time series patches: during pretraining, it dynamically samples visible context and masked target patches, exposing the model to diverse context-target configurations without sequential decoding. At inference time, the same pretrained model supports one-step forecasting, confidence-based iterative refinement, and next-token-style generation through a unified mask-and-predict interface, enabling controlled comparison of inference schedules under identical architecture, data, and objective. Evaluated on BLAST and multiple zero-shot benchmarks, FlexCast achieves accuracy competitive with or better than strong TSFM baselines while substantially reducing inference latency. Ablations confirm that context diversity and contiguous target masking drive performance, while left-to-right decoding does not yield consistent gains once flexible context-target relationships have been learned, suggesting that the key benefit of AR-style TSFMs lies in flexible training rather than sequential inference. The code is available at https://github.com/BenchCouncil/FlexCast.

Figures & tables

Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Instance-Adaptive Prompts as Context for Time-Series Foundation Models

    Sep 28, 2026Zehao Xiao, Shifeng Xie, Lei Zan +5Time Series ForecastingPrompt Tuning

  2. Retrieval Mechanisms Surpass Long-Context Scaling in Time Series Forecasting

    May 6, 2026Rishi Ahuja, Kumar Prateek, Simranjit Singh +1Time Series ForecastingLong-Context Modeling

  3. Latent Inference-Time Guidance of Time Series Foundation Models

    Sep 29, 2026Chloé Hashimoto-Cullen, Amaury Durand, Laurent Bozzi +3Time Series ForecastingInference-Time Guidance