What question did this study set out to answer?

This research investigates whether the architecture of neural operators should correspond to the strength of multiphysics coupling they model.

July 5, 2026Communications EngineeringOpen Access

Network architecture follows coupling in multiphysics systems: single vs. multiple branches in DeepONet and S-DeepONet

View Full Paper

Ask AI

Bookmark

Authors

JPJaewan ParkUniversity of Illinois Urbana-Champaign KKKazuma KobayashiUniversity of Illinois Urbana-Champaign QLQibang LiuGeorgia Institute of Technology

Discussion

Member takes

Implication

Randomized trial compares single-branch and multi-branch neural operators for predicting tightly and loosely coupled systems, highlighting performance variations.

Key Points

This research investigates whether the architecture of neural operators should correspond to the strength of multiphysics coupling they model.
Compared single-branch and multi-branch neural networks in feedforward and sequential forms across three physical systems.
Evaluated performance in a reaction-diffusion problem, a thermo-electrical problem, and a viscoplastic thermo-mechanical model.
Single-branch networks achieved higher accuracy in tightly coupled scenarios compared to multi-branch variants.
Multi-branch networks performed better in decoupled or single-physics cases.
Surrogates provided predictions up to 1.8 × 10^4 times faster than traditional physics-based solvers.

Cite This Study

Park et al. (2026) studied this question.

synapsesocial.com/papers/6a49f70df5d1d45b288011ad https://doi.org/10.1038/s44172-026-00714-4

View Full Paper

Ask AI

Bookmark

July 5, 2026Communications EngineeringOpen Access

Network architecture follows coupling in multiphysics systems: single vs. multiple branches in DeepONet and S-DeepONet

View Full Paper

Ask AI

Bookmark

Authors

JPJaewan ParkUniversity of Illinois Urbana-Champaign KKKazuma KobayashiUniversity of Illinois Urbana-Champaign QLQibang LiuGeorgia Institute of Technology

Discussion

Member takes

Implication

Randomized trial compares single-branch and multi-branch neural operators for predicting tightly and loosely coupled systems, highlighting performance variations.

Key Points

This research investigates whether the architecture of neural operators should correspond to the strength of multiphysics coupling they model.
Compared single-branch and multi-branch neural networks in feedforward and sequential forms across three physical systems.
Evaluated performance in a reaction-diffusion problem, a thermo-electrical problem, and a viscoplastic thermo-mechanical model.
Single-branch networks achieved higher accuracy in tightly coupled scenarios compared to multi-branch variants.
Multi-branch networks performed better in decoupled or single-physics cases.
Surrogates provided predictions up to 1.8 × 10^4 times faster than traditional physics-based solvers.

Cite This Study

Park et al. (2026) studied this question.

synapsesocial.com/papers/6a49f70df5d1d45b288011ad https://doi.org/10.1038/s44172-026-00714-4

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark