Learning Deep Time-index Models for Time Series Forecasting

Gerald Woo, Chenghao Liu, Doyen Sahoo, Akshat Kumar, Steven Hoi
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:37217-37237, 2023.

Abstract

Deep learning has been actively applied to time series forecasting, leading to a deluge of new methods, belonging to the class of historical-value models. Yet, despite the attractive properties of time-index models, such as being able to model the continuous nature of underlying time series dynamics, little attention has been given to them. Indeed, while naive deep time-index models are far more expressive than the manually predefined function representations of classical time-index models, they are inadequate for forecasting, being unable to generalize to unseen time steps due to the lack of inductive bias. In this paper, we propose DeepTime, a meta-optimization framework to learn deep time-index models which overcome these limitations, yielding an efficient and accurate forecasting model. Extensive experiments on real world datasets in the long sequence time-series forecasting setting demonstrate that our approach achieves competitive results with state-of-the-art methods, and is highly efficient. Code is available at https://github.com/salesforce/DeepTime.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-woo23b, title = {Learning Deep Time-index Models for Time Series Forecasting}, author = {Woo, Gerald and Liu, Chenghao and Sahoo, Doyen and Kumar, Akshat and Hoi, Steven}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {37217--37237}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/woo23b/woo23b.pdf}, url = {https://proceedings.mlr.press/v202/woo23b.html}, abstract = {Deep learning has been actively applied to time series forecasting, leading to a deluge of new methods, belonging to the class of historical-value models. Yet, despite the attractive properties of time-index models, such as being able to model the continuous nature of underlying time series dynamics, little attention has been given to them. Indeed, while naive deep time-index models are far more expressive than the manually predefined function representations of classical time-index models, they are inadequate for forecasting, being unable to generalize to unseen time steps due to the lack of inductive bias. In this paper, we propose DeepTime, a meta-optimization framework to learn deep time-index models which overcome these limitations, yielding an efficient and accurate forecasting model. Extensive experiments on real world datasets in the long sequence time-series forecasting setting demonstrate that our approach achieves competitive results with state-of-the-art methods, and is highly efficient. Code is available at https://github.com/salesforce/DeepTime.} }
Endnote
%0 Conference Paper %T Learning Deep Time-index Models for Time Series Forecasting %A Gerald Woo %A Chenghao Liu %A Doyen Sahoo %A Akshat Kumar %A Steven Hoi %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-woo23b %I PMLR %P 37217--37237 %U https://proceedings.mlr.press/v202/woo23b.html %V 202 %X Deep learning has been actively applied to time series forecasting, leading to a deluge of new methods, belonging to the class of historical-value models. Yet, despite the attractive properties of time-index models, such as being able to model the continuous nature of underlying time series dynamics, little attention has been given to them. Indeed, while naive deep time-index models are far more expressive than the manually predefined function representations of classical time-index models, they are inadequate for forecasting, being unable to generalize to unseen time steps due to the lack of inductive bias. In this paper, we propose DeepTime, a meta-optimization framework to learn deep time-index models which overcome these limitations, yielding an efficient and accurate forecasting model. Extensive experiments on real world datasets in the long sequence time-series forecasting setting demonstrate that our approach achieves competitive results with state-of-the-art methods, and is highly efficient. Code is available at https://github.com/salesforce/DeepTime.
APA
Woo, G., Liu, C., Sahoo, D., Kumar, A. & Hoi, S.. (2023). Learning Deep Time-index Models for Time Series Forecasting. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:37217-37237 Available from https://proceedings.mlr.press/v202/woo23b.html.

Related Material