TANTE: Time-Adaptive Operator Learning via Neural Taylor Expansion
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909826438660096 |
|---|---|
| author | Wu, Zhikai Wang, Sifan Zhang, Shiyang He, Sizhuang Zhu, Min Jiao, Anran Lu, Lu van Dijk, David |
| author_facet | Wu, Zhikai Wang, Sifan Zhang, Shiyang He, Sizhuang Zhu, Min Jiao, Anran Lu, Lu van Dijk, David |
| contents | Operator learning for time-dependent partial differential equations (PDEs) has seen rapid progress in recent years, enabling efficient approximation of complex spatiotemporal dynamics. However, most existing methods rely on fixed time step sizes during rollout, which limits their ability to adapt to varying temporal complexity and often leads to error accumulation. Here, we propose the Time-Adaptive Transformer with Neural Taylor Expansion (TANTE), a novel operator-learning framework that produces continuous-time predictions with adaptive step sizes. TANTE predicts future states by performing a Taylor expansion at the current state, where neural networks learn both the higher-order temporal derivatives and the local radius of convergence. This allows the model to dynamically adjust its rollout based on the local behavior of the solution, thereby reducing cumulative error and improving computational efficiency. We demonstrate the effectiveness of TANTE across a wide range of PDE benchmarks, achieving superior accuracy and adaptability compared to fixed-step baselines, delivering accuracy gains of 60-80 % and speed-ups of 30-40 % at inference time. The code is publicly available at https://github.com/zwu88/TANTE for transparency and reproducibility. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_08574 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | TANTE: Time-Adaptive Operator Learning via Neural Taylor Expansion Wu, Zhikai Wang, Sifan Zhang, Shiyang He, Sizhuang Zhu, Min Jiao, Anran Lu, Lu van Dijk, David Machine Learning Artificial Intelligence Operator learning for time-dependent partial differential equations (PDEs) has seen rapid progress in recent years, enabling efficient approximation of complex spatiotemporal dynamics. However, most existing methods rely on fixed time step sizes during rollout, which limits their ability to adapt to varying temporal complexity and often leads to error accumulation. Here, we propose the Time-Adaptive Transformer with Neural Taylor Expansion (TANTE), a novel operator-learning framework that produces continuous-time predictions with adaptive step sizes. TANTE predicts future states by performing a Taylor expansion at the current state, where neural networks learn both the higher-order temporal derivatives and the local radius of convergence. This allows the model to dynamically adjust its rollout based on the local behavior of the solution, thereby reducing cumulative error and improving computational efficiency. We demonstrate the effectiveness of TANTE across a wide range of PDE benchmarks, achieving superior accuracy and adaptability compared to fixed-step baselines, delivering accuracy gains of 60-80 % and speed-ups of 30-40 % at inference time. The code is publicly available at https://github.com/zwu88/TANTE for transparency and reproducibility. |
| title | TANTE: Time-Adaptive Operator Learning via Neural Taylor Expansion |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2502.08574 |