论文

FlexCast 提出灵活上下文学习框架,挑战时间序列自回归基础模型范式

FlexCast: Flexible Context Learning for Time Series Foundation Models

精选理由

这篇论文把自回归时间序列模型的优势拆开验证了:收益主要来自灵活训练而非逐步解码。做时序预测的开发者可以用掩码式推理拿到相近精度和更低延迟,代码也开源了,可直接对照复现。

arXiv 论文提出 FlexCast,将时间序列预测重构为对时间序列 patch 的可变长度掩码预测,在预训练阶段动态采样可见上下文与被掩码的目标 patch,使模型在无需序列解码的情况下接触多样的上下文-目标配置。

正文

同一预训练模型通过统一的掩码-预测接口支持三种推理方式:一步预测、基于置信度的迭代精化、以及 next-token 式生成,从而在相同架构、数据和目标下可控地比较不同推理调度。在 BLAST 及多个 zero-shot 基准上,FlexCast 取得与强 TSFM 基线相当或更好的准确率,同时大幅降低推理延迟。

消融实验显示,上下文多样性和连续目标掩码是性能来源,而左到右解码在学习了灵活上下文-目标关系后不再带来一致收益,说明自回归式 TSFM 的关键在于灵活训练而非序列推理。代码已在 GitHub 开源(BenchCouncil/FlexCast)。

原文 · arXiv cs.AI

FlexCast: Flexible Context Learning for Time Series Foundation Models

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.

产业坐标与信源核验

本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:FlexCast、时间序列预测、基础模型、掩码预测,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。

信源核验说明:本文原始线索采编自「arXiv cs.AI」(@Zuyang Shen, Luqi Gong, Fanda Fan, Bingsong Wang, Xiang Ao, Yahui Wei, Minghao Liu, Chunjie Luo, Jianfeng Zhan),经 AITOP 多维评分管线结构化提炼。