arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2505.02417cs.LGcs.AI
收藏

T2S:基于文本到序列扩散模型的高分辨率时间序列生成

T2S: High-resolution Time Series Generation with Text-to-Series Diffusion Models

Xidian University(西安电子科技大学);Griffith University(格里菲斯大学);Yunnan University(云南大学);Carnegie Mellon University(卡内基梅隆大学);Zhejiang University(浙江大学);Augusta University(奥古斯塔大学);The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州))

Yunfeng Ge, Jiawei Li, Yiji Zhao, Haomin Wen, Zhao Li, Meikang Qiu, Hongyan Li, Ming Jin, Shirui Pan

AI总结:

本文提出T2S扩散框架,通过长度自适应变分自编码器和Flow Matching对齐文本与时间序列,实现任意长度的高分辨率时间序列生成,并在13个数据集上取得最优性能。

摘要:

文本到时间序列生成在解决数据稀疏、数据不平衡以及跨领域多模态时间序列数据集可用性有限等挑战方面具有重要潜力。尽管扩散模型在文本到X(例如视觉和音频数据)生成方面取得了显著成功,但其在时间序列生成中的应用仍处于起步阶段。现有方法面临两个关键局限:(1)缺乏对通用时间序列文本描述的系统的探索,现有描述往往针对特定领域,难以泛化;(2)无法生成任意长度的时间序列,限制了其在现实场景中的适用性。在本工作中,我们首先将时间序列文本描述划分为三个层级:点级、片段级和实例级。此外,我们引入了一个新的片段级数据集,包含超过600,000个高分辨率时间序列-文本对。其次,我们提出了Text-to-Series(T2S),这是一个基于扩散的框架,以领域无关的方式弥合自然语言与时间序列之间的差距。T2S采用长度自适应变分自编码器将不同长度的时间序列编码为一致的潜在嵌入。在此基础上,T2S利用Flow Matching并将Diffusion Transformer作为去噪器,有效地将文本表示与潜在嵌入对齐。我们在多个长度上以交错范式训练T2S,使其能够生成任意所需长度的序列。广泛的评估表明,T2S在跨越12个领域的13个数据集上取得了最先进的性能。

英文摘要:

Text-to-Time Series generation holds significant potential to address challenges such as data sparsity, imbalance, and limited availability of multimodal time series datasets across domains. While diffusion models have achieved remarkable success in Text-to-X (e.g., vision and audio data) generation, their use in time series generation remains in its nascent stages. Existing approaches face two critical limitations: (1) the lack of systematic exploration of general-proposed time series captions, which are often domain-specific and struggle with generalization; and (2) the inability to generate time series of arbitrary lengths, limiting their applicability to real-world scenarios. In this work, we first categorize time series captions into three levels: point-level, fragment-level, and instance-level. Additionally, we introduce a new fragment-level dataset containing over 600,000 high-resolution time series-text pairs. Second, we propose Text-to-Series (T2S), a diffusion-based framework that bridges the gap between natural language and time series in a domain-agnostic manner. T2S employs a length-adaptive variational autoencoder to encode time series of varying lengths into consistent latent embeddings. On top of that, T2S effectively aligns textual representations with latent embeddings by utilizing Flow Matching and employing Diffusion Transformer as the denoiser. We train T2S in an interleaved paradigm across multiple lengths, allowing it to generate sequences of any desired length. Extensive evaluations demonstrate that T2S achieves state-of-the-art performance across 13 datasets spanning 12 domains.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑