Machine learningTime-series forecasting
Moirai：通用时间序列预测Transformer
Moirai 是 Salesforce Research 的 Gerald Woo 及其同事于 2024 年推出并提交给 ICML 的通用时间序列预测基础模型。其核心思想是在一个极其多样化的时间序列数据语料库（LOTSA）上预训练单个大型Transformer，该语料库涵盖了多种领域和频率，从而能够在零样本（zero-shot）和少样本（few-shot）场景下对未见过的数据集进行预测，而无需进行特定任务的再训练。Moirai 采用了基于块（patch-based）的标记化、任意变量（any-variate）注意力以及混合分布（mixture-of-distributions）输出头，以在统一的架构中处理可变频率、多变量和概率预测。
阅读完整方法
仅限会员
登录使用免费账户登录即可阅读本节。
Method map
The neighbourhood of related methods — select a node to explore.
来源
- Woo, G., Liu, C., Kumar, A., Xiong, C., Savarese, S., & Sahoo, D. (2024). Unified training of universal time series forecasting transformers. ICML. link ↗
如何引用本页
ScholarGate. (2026, June 2). Moirai (Universal Time-Series Forecasting Transformer). ScholarGate. https://scholargate.app/zh/deep-learning/moirai
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
Compare side by side →