LETS Forecast: Learning Embedology for Time Series Forecasting

Abrar Majeedi, Viswanatha Reddy Gajjala, Satya Sai Srinath Namburi Gnvv, Nada Magdi Elkordi, Yin Li
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:42668-42690, 2025.

Abstract

Real-world time series are often governed by complex nonlinear dynamics. Understanding these underlying dynamics is crucial for precise future prediction. While deep learning has achieved major success in time series forecasting, many existing approaches do not explicitly model the dynamics. To bridge this gap, we introduce DeepEDM, a framework that integrates nonlinear dynamical systems modeling with deep neural networks. Inspired by empirical dynamic modeling (EDM) and rooted in Takens’ theorem, DeepEDM presents a novel deep model that learns a latent space from time-delayed embeddings, and employs kernel regression to approximate the underlying dynamics, while leveraging efficient implementation of softmax attention and allowing for accurate prediction of future time steps. To evaluate our method, we conduct comprehensive experiments on synthetic data of nonlinear dynamical systems as well as real-world time series across domains. Our results show that DeepEDM is robust to input noise, and outperforms state-of-the-art methods in forecasting accuracy. Our code is available at: https://abrarmajeedi.github.io/deep_edm.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-majeedi25a, title = {{LETS} Forecast: Learning Embedology for Time Series Forecasting}, author = {Majeedi, Abrar and Gajjala, Viswanatha Reddy and Namburi Gnvv, Satya Sai Srinath and Elkordi, Nada Magdi and Li, Yin}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {42668--42690}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/majeedi25a/majeedi25a.pdf}, url = {https://proceedings.mlr.press/v267/majeedi25a.html}, abstract = {Real-world time series are often governed by complex nonlinear dynamics. Understanding these underlying dynamics is crucial for precise future prediction. While deep learning has achieved major success in time series forecasting, many existing approaches do not explicitly model the dynamics. To bridge this gap, we introduce DeepEDM, a framework that integrates nonlinear dynamical systems modeling with deep neural networks. Inspired by empirical dynamic modeling (EDM) and rooted in Takens’ theorem, DeepEDM presents a novel deep model that learns a latent space from time-delayed embeddings, and employs kernel regression to approximate the underlying dynamics, while leveraging efficient implementation of softmax attention and allowing for accurate prediction of future time steps. To evaluate our method, we conduct comprehensive experiments on synthetic data of nonlinear dynamical systems as well as real-world time series across domains. Our results show that DeepEDM is robust to input noise, and outperforms state-of-the-art methods in forecasting accuracy. Our code is available at: https://abrarmajeedi.github.io/deep_edm.} }
Endnote
%0 Conference Paper %T LETS Forecast: Learning Embedology for Time Series Forecasting %A Abrar Majeedi %A Viswanatha Reddy Gajjala %A Satya Sai Srinath Namburi Gnvv %A Nada Magdi Elkordi %A Yin Li %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-majeedi25a %I PMLR %P 42668--42690 %U https://proceedings.mlr.press/v267/majeedi25a.html %V 267 %X Real-world time series are often governed by complex nonlinear dynamics. Understanding these underlying dynamics is crucial for precise future prediction. While deep learning has achieved major success in time series forecasting, many existing approaches do not explicitly model the dynamics. To bridge this gap, we introduce DeepEDM, a framework that integrates nonlinear dynamical systems modeling with deep neural networks. Inspired by empirical dynamic modeling (EDM) and rooted in Takens’ theorem, DeepEDM presents a novel deep model that learns a latent space from time-delayed embeddings, and employs kernel regression to approximate the underlying dynamics, while leveraging efficient implementation of softmax attention and allowing for accurate prediction of future time steps. To evaluate our method, we conduct comprehensive experiments on synthetic data of nonlinear dynamical systems as well as real-world time series across domains. Our results show that DeepEDM is robust to input noise, and outperforms state-of-the-art methods in forecasting accuracy. Our code is available at: https://abrarmajeedi.github.io/deep_edm.
APA
Majeedi, A., Gajjala, V.R., Namburi Gnvv, S.S.S., Elkordi, N.M. & Li, Y.. (2025). LETS Forecast: Learning Embedology for Time Series Forecasting. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:42668-42690 Available from https://proceedings.mlr.press/v267/majeedi25a.html.

Related Material