AdamFLIP: Adaptive Momentum Feedback Linearization Optimization for Hard Constrained PINN Training
Binghang Lu, Runyu Zhang, Changhong Mou, Guang Lin · 2026
Read AdamFLIP: Adaptive Momentum Feedback Linearization Optimization for Hard Constrained PINN Training by Binghang Lu, Runyu Zhang, Changhong Mou, Guang Lin (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-08408.txt.
Description
Binghang Lu, Runyu Zhang, Changhong Mou, Na Li, and coauthors enforce the physics in a neural PDE solver instead of merely penalizing violations. Standard physics-informed networks mix residuals, data mismatch, and boundary losses with hand-chosen weights, which often yields ill-conditioning and constraints satisfied only loosely.
AdamFLIP recasts training as an equality-constrained optimization problem. Constraint residuals are treated as the output of a controlled system, and the Lagrange multiplier becomes a feedback signal that drives them toward stable linear contraction. Adam-style moment adaptation then tunes that feedback. The hard-constraint framing aims at solutions that respect the governing equations by construction rather than by bargaining among loss weights.
Catalog record
| CREATOR | Binghang Lu, Runyu Zhang, Changhong Mou, Guang Lin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 112 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.08408