Metraipolis › archive › dc_d2d450ecf41c
Task-Contribution Dynamics of Adam in Physics-Informed Neural Networks: Slow-Manifold Analysis and Gradient Modulation — arXiv:2611.08910
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.08910
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-09T22:41:53.655Z
- SHA-256 of the archived text
43adb691adbee0efab2036309ae96c4744d8fae9c4f519c917f1e57c4075d585- Stable URL of this page
https://metraipolis.com/documento/dc_d2d450ecf41c
How to cite
Task-Contribution Dynamics of Adam in Physics-Informed Neural Networks: Slow-Manifold Analysis and Gradient Modulation — arXiv:2611.08910 — arXiv (en), https://arxiv.org/abs/2611.08910 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d2d450ecf41c (Metraipolis archive dc_d2d450ecf41c, sha256 43adb691adbee0efab2036309ae96c4744d8fae9c4f519c917f1e57c4075d585)
BibTeX
@misc{dcd2d450ecf41c,
title = {Task-Contribution Dynamics of Adam in Physics-Informed Neural Networks: Slow-Manifold Analysis and Gradient Modulation — arXiv:2611.08910},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.08910},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d2d450ecf41c; sha256 43adb691adbee0efab2036309ae96c4744d8fae9c4f519c917f1e57c4075d585},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.08910. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Physics-informed neural networks (PINNs) often suffer from imbalanced optimization of competing loss components, particularly when approximating solutions with high-frequency or multiscale features. Although existing balancing strategies have proven effective in alleviating this difficulty, the mechanisms governing the evolution of individual task contributions to Adam's parameter updates remain insufficiently understood. We introduce task-wise indicators to characterize this evolution and identify three distinct training regimes: smooth variation, intermittent spikes, and persistent oscillations. Focusing on the early smooth regime, we combine discrete-time analysis with a continuous-time singular perturbation framework under the full-batch setting and the parameter regime $(1-\beta_2)/(1-\beta_1)\ll1$. A first-order approximation of the locally attracting slow manifold reveals two factors governing the step-to-step redistribution of task contributions: curvature-related alignment disparities and differences in the coordinatewise distributions of preconditioned task gradients. Both factors involve Adam's shared second-moment estimate. Inspired by AdaTask, we propose Gradient-Modulated Adam (GMAdam), which modulates incoming task gradients using task-specific second-moment estimates. Numerical experiments on high-frequency Poisson and wave equations, the Allen-Cahn equation, and two-dimensional lid-driven cavity flow ($Re=100,400$) demonstrate consistent improvements in predictive accuracy over AdaTask at comparable computational cost. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Numerical Analysis (math.NA) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.08910 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-09T22:33:47.173Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).