Rethinking Time Encoding via Learnable Transformation Functions

Xi Chen, Yateng Tang, Jiarong Xu, Jiawei Zhang, Siwei Zhang, Sijia Peng, Xuehao Zheng, Yun Xiong
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:9073-9104, 2025.

Abstract

Effectively modeling time information and incorporating it into applications or models involving chronologically occurring events is crucial. Real-world scenarios often involve diverse and complex time patterns, which pose significant challenges for time encoding methods. While previous methods focus on capturing time patterns, many rely on specific inductive biases, such as using trigonometric functions to model periodicity. This narrow focus on single-pattern modeling makes them less effective in handling the diversity and complexities of real-world time patterns. In this paper, we investigate to improve the existing commonly used time encoding methods and introduce Learnable Transformation-based Generalized Time Encoding (LeTE). We propose using deep function learning techniques to parameterize nonlinear transformations in time encoding, making them learnable and capable of modeling generalized time patterns, including diverse and complex temporal dynamics. By enabling learnable transformations, LeTE encompasses previous methods as specific cases and allows seamless integration into a wide range of tasks. Through extensive experiments across diverse domains, we demonstrate the versatility and effectiveness of LeTE.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-chen25bh, title = {Rethinking Time Encoding via Learnable Transformation Functions}, author = {Chen, Xi and Tang, Yateng and Xu, Jiarong and Zhang, Jiawei and Zhang, Siwei and Peng, Sijia and Zheng, Xuehao and Xiong, Yun}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {9073--9104}, 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/chen25bh/chen25bh.pdf}, url = {https://proceedings.mlr.press/v267/chen25bh.html}, abstract = {Effectively modeling time information and incorporating it into applications or models involving chronologically occurring events is crucial. Real-world scenarios often involve diverse and complex time patterns, which pose significant challenges for time encoding methods. While previous methods focus on capturing time patterns, many rely on specific inductive biases, such as using trigonometric functions to model periodicity. This narrow focus on single-pattern modeling makes them less effective in handling the diversity and complexities of real-world time patterns. In this paper, we investigate to improve the existing commonly used time encoding methods and introduce Learnable Transformation-based Generalized Time Encoding (LeTE). We propose using deep function learning techniques to parameterize nonlinear transformations in time encoding, making them learnable and capable of modeling generalized time patterns, including diverse and complex temporal dynamics. By enabling learnable transformations, LeTE encompasses previous methods as specific cases and allows seamless integration into a wide range of tasks. Through extensive experiments across diverse domains, we demonstrate the versatility and effectiveness of LeTE.} }
Endnote
%0 Conference Paper %T Rethinking Time Encoding via Learnable Transformation Functions %A Xi Chen %A Yateng Tang %A Jiarong Xu %A Jiawei Zhang %A Siwei Zhang %A Sijia Peng %A Xuehao Zheng %A Yun Xiong %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-chen25bh %I PMLR %P 9073--9104 %U https://proceedings.mlr.press/v267/chen25bh.html %V 267 %X Effectively modeling time information and incorporating it into applications or models involving chronologically occurring events is crucial. Real-world scenarios often involve diverse and complex time patterns, which pose significant challenges for time encoding methods. While previous methods focus on capturing time patterns, many rely on specific inductive biases, such as using trigonometric functions to model periodicity. This narrow focus on single-pattern modeling makes them less effective in handling the diversity and complexities of real-world time patterns. In this paper, we investigate to improve the existing commonly used time encoding methods and introduce Learnable Transformation-based Generalized Time Encoding (LeTE). We propose using deep function learning techniques to parameterize nonlinear transformations in time encoding, making them learnable and capable of modeling generalized time patterns, including diverse and complex temporal dynamics. By enabling learnable transformations, LeTE encompasses previous methods as specific cases and allows seamless integration into a wide range of tasks. Through extensive experiments across diverse domains, we demonstrate the versatility and effectiveness of LeTE.
APA
Chen, X., Tang, Y., Xu, J., Zhang, J., Zhang, S., Peng, S., Zheng, X. & Xiong, Y.. (2025). Rethinking Time Encoding via Learnable Transformation Functions. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:9073-9104 Available from https://proceedings.mlr.press/v267/chen25bh.html.

Related Material