Saved in:
Bibliographic Details
Main Authors: Ye, Zaijun, Zhang, Chen-Song, Wang, Wansheng
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.20721
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915306238115840
author Ye, Zaijun
Zhang, Chen-Song
Wang, Wansheng
author_facet Ye, Zaijun
Zhang, Chen-Song
Wang, Wansheng
contents Neural operators have emerged as powerful tools for learning solution operators of partial differential equations. However, in time-dependent problems, standard training strategies such as teacher forcing introduce a mismatch between training and inference, leading to compounding errors in long-term autoregressive predictions. To address this issue, we propose Recurrent Neural Operators (RNOs)-a novel framework that integrates recurrent training into neural operator architectures. Instead of conditioning each training step on ground-truth inputs, RNOs recursively apply the operator to their own predictions over a temporal window, effectively simulating inference-time dynamics during training. This alignment mitigates exposure bias and enhances robustness to error accumulation. Theoretically, we show that recurrent training can reduce the worst-case exponential error growth typical of teacher forcing to linear growth. Empirically, we demonstrate that recurrently trained Multigrid Neural Operators significantly outperform their teacher-forced counterparts in long-term accuracy and stability on standard benchmarks. Our results underscore the importance of aligning training with inference dynamics for robust temporal generalization in neural operator learning.
format Preprint
id arxiv_https___arxiv_org_abs_2505_20721
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Recurrent Neural Operators: Stable Long-Term PDE Prediction
Ye, Zaijun
Zhang, Chen-Song
Wang, Wansheng
Machine Learning
Numerical Analysis
65M70, 68T07, 68U20
Neural operators have emerged as powerful tools for learning solution operators of partial differential equations. However, in time-dependent problems, standard training strategies such as teacher forcing introduce a mismatch between training and inference, leading to compounding errors in long-term autoregressive predictions. To address this issue, we propose Recurrent Neural Operators (RNOs)-a novel framework that integrates recurrent training into neural operator architectures. Instead of conditioning each training step on ground-truth inputs, RNOs recursively apply the operator to their own predictions over a temporal window, effectively simulating inference-time dynamics during training. This alignment mitigates exposure bias and enhances robustness to error accumulation. Theoretically, we show that recurrent training can reduce the worst-case exponential error growth typical of teacher forcing to linear growth. Empirically, we demonstrate that recurrently trained Multigrid Neural Operators significantly outperform their teacher-forced counterparts in long-term accuracy and stability on standard benchmarks. Our results underscore the importance of aligning training with inference dynamics for robust temporal generalization in neural operator learning.
title Recurrent Neural Operators: Stable Long-Term PDE Prediction
topic Machine Learning
Numerical Analysis
65M70, 68T07, 68U20
url https://arxiv.org/abs/2505.20721