Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
February 2, 2026EntropyOpen Access

QEKI: A Quantum–Classical Framework for Efficient Bayesian Inversion of PDEs

View Full Paper
Ask AI
Bookmark
Share

Authors

JYJiawei YongShanghaiTech UniversitySTSihai TangUniversity of North Texas

Discussion

Loading...

Member takes

Implication

This framework demonstrates improved Bayesian inversion in PDEs, indicating a hybrid quantum-classical approach for enhanced computational efficiency.

Key Points

  • The aim is to develop an efficient framework for solving Bayesian inverse problems in the context of partial differential equations (PDEs).
  • Introduced Quantum-Encodable Bayesian PINNs (QEKI) for training Bayesian PINNs.
  • Paired Quantum Neural Networks (QNNs) with Ensemble Kalman Inversion (EKI) for efficient sampling.
  • Utilized a gradient-free approach to address quantum optimization issues, particularly the barren plateau.
  • QEKI achieved precise inversions for 1D and 2D nonlinear PDEs.
  • Demonstrated substantial parameter compression compared to classical networks.
  • Proven effective even in noisy environments.

Cite This Study

Yong et al. (2026) studied this question.

synapsesocial.com/papers/6980fe35c1c9540dea81019dhttps://doi.org/10.3390/e28020156
View Full Paper
Ask AI
Bookmark
Share

Also Consider

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

  1. 1Physics-informed quantum neural network for solving forward and inverse problems of partial differential equations2024 · 21 citations
  2. 2Quantum Physics-Informed Neural Networks for Maxwell’s Equations: Circuit Design, “Black Hole” Barren Plateaus Mitigation, and GPU Acceleration2025
  3. 3Quantum physics-informed neural networks for Maxwell’s equations: circuit design, “black hole” barren plateaus mitigation, and GPU acceleration2026
  4. 4Quantum Physics-Informed Neural Network for Scattered Wavefields2026
  5. 5Solving Differential Equation with Quantum-Circuit Enhanced Physics-Informed Neural Networks2025