arXiv：cs.LG· Haoran Zhang, Haixuan Liu, Yong Liu, Yunzhong Qiu, Yuxuan Wang, Jianmin Wang, Mingsheng Long·· 4 小时前AI 评分32
DiTS：多模态扩散 Transformer 如何成为时间序列预测器
DiTS: Multimodal Diffusion Transformers Are Time Series Forecasters
AI 导读
研究者提出 DiTS，一种面向协变量感知预测的多模态扩散 Transformer，将内生目标与外生协变量建模为不同模态，并联合条件化未来生成。DiTS 引入 Time-aligned Modulation，在 AdaLN 基础上利用时间对齐先验，从对齐协变量与扩散时间生成 patch 级调制参数，配合 flow matching 使协变量依赖的分布信息参与速度预测。
正文
Abstract:While generative modeling facilitates probabilistic time series forecasting, incorporating heterogeneous exogenous information remains challenging. Diffusion Transformers (DiT) provide a scalable framework for conditional generation, yet their adaptation to forecasting calls for conditioning mechanisms tailored to time series. Endogenous targets and exogenous covariates differ in sources, semantics, and statistical characteristics, while sharing temporal coordinates that support fine-grained conditional guidance. Covariates can describe future variability and temporal dependence beyond the conditional mean targeted by direct regression. Motivated by these considerations, we propose Diffusion Transformers for Time Series (DiTS), a Multimodal Diffusion Transformer for covariate-aware forecasting. DiTS models endogenous targets and exogenous covariates as distinct modalities, jointly conditioning future generation on target history and available covariates. Flow matching makes covariate-dependent distributional information relevant to velocity prediction conditioned on noisy future states. We introduce Time-aligned Modulation, extending AdaLN with the temporal-alignment prior to generate patch-wise modulation parameters from aligned covariates and diffusion time. Across diverse covariate-aware forecasting tasks, DiTS achieves strong performance in both deterministic and probabilistic forecasting, demonstrating the effectiveness of conditional generation for both point and distributional forecasting.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2602.06597 [cs.LG] |
| (or arXiv:2602.06597v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2602.06597 arXiv-issued DOI via DataCite |
Submission history
From: Haoran Zhang [view email]
[v1]
Fri, 6 Feb 2026 10:48:13 UTC (6,753 KB)
[v2]
Wed, 7 Oct 2026 15:43:43 UTC (1,336 KB)
来源：arXiv：cs.LG · arxiv.org