arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.08424stat.MLcs.LG
收藏

使用 TiDE 进行长期预测:Time-series Dense Encoder(时间序列密集编码器)

Long-term Forecasting with TiDE: Time-series Dense Encoder

Google Research(谷歌研究院);Google Cloud(谷歌云);University of California, San Diego(加利福尼亚大学圣迭戈分校)

Abhimanyu Das, Weihao Kong, Andrew Leach, Shaan Mathur, Rajat Sen, Rose Yu

AI总结:

本文提出基于 MLP 的编码器-解码器 TiDE,用于长期时间序列预测,可处理协变量与非线性依赖,并在理论上证明线性变体近最优,实验上性能匹敌或超越既有方法且速度快 5-10 倍。

摘要:

近期研究表明,在长期时间序列预测中,简单线性模型的表现可以优于若干基于 Transformer 的方法。受此启发,我们提出了一种基于多层感知机(MLP)的编码器-解码器模型 Time-series Dense Encoder(TiDE,时间序列密集编码器),用于长期时间序列预测;该模型兼具线性模型的简洁性与速度,同时能够处理协变量和非线性依赖关系。在理论方面,我们证明了在某些假设下,我们模型最简单的线性类似物在线性动力系统(LDS)上能够达到接近最优的错误率。在实验方面,我们表明,在流行的长期时间序列预测基准上,我们的方法能够匹敌或优于以往方法,同时速度比最佳的基于 Transformer 的模型快 5 至 10 倍。

英文摘要:

Recent work has shown that simple linear models can outperform several Transformer based approaches in long term time-series forecasting. Motivated by this, we propose a Multi-layer Perceptron (MLP) based encoder-decoder model, Time-series Dense Encoder (TiDE), for long-term time-series forecasting that enjoys the simplicity and speed of linear models while also being able to handle covariates and non-linear dependencies. Theoretically, we prove that the simplest linear analogue of our model can achieve near optimal error rate for linear dynamical systems (LDS) under some assumptions. Empirically, we show that our method can match or outperform prior approaches on popular long-term time-series forecasting benchmarks while being 5-10x faster than the best Transformer based model.

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。

↑