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

By Abhimanyu Das, Weihao Kong, Andrew B. Leach, Shaan Mathur, Rajat Sen, Rose Yu. Trans. Mach. Learn. Res., 2023

600 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Statistical and Nonlinear Physics, Signal Processing

View Paper: https://www.semanticscholar.org/paper/2ced138789c8a1c39f0f57c8fbb18b94e6ed8034

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Haixu Wu
2. Shiyu Wang
3. Mingsheng Long
4. Jianmin Wang
5. Kate Saenko
6. Zhangjie Cao
7. Jian Liang
8. Jiashi Feng

In Statistical and Nonlinear Physics

1. Haixu Wu
2. Mingsheng Long

In Signal Processing

1. Shiyu Wang
2. Lintao Ma
3. Haixu Wu
4. Tengge Hu
5. Mingsheng Long

Citations in the corpus

• Is Mamba Effective for Time Series Forecasting? (2024, neutral): “Consequently, there are weak VCs between these variates, and the employment of Mamba VC Encoding layer by S-Mamba can’t bring useful information and even may inadvertently introduce noise into the predictive model, thus impeding its predictive accuracy.”
• iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (2023, methodology): “…our efficient training strategy and iTransformer with the efficient flow attention module (Wu et al., 2022); linear models: DLinear (Zeng et al., 2023) and TiDE (Das et al., 2023); Transformers: Transformer (Vaswani et al., 2017), PatchTST (Nie et al., 2023), and Crossformer (Zhang & Yan, 2023).”