.png){kind=icon}
Return
Differentiable singular value decomposition (SVD)
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ymssp.2025.112817.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Singular value decomposition (SVD) is widely used in modal analysis, such as proper orthogonal decomposition (POD) and resolvent analysis, to extract key features from imaginary problems. SVD derivatives need to be computed efficiently to enable the large-scale design optimization. However, for a general imaginary matrix, no method can accurately compute this derivative to machine precision and remain scalable with respect to the number of design variables without requiring the all of the singular variables. We propose two algorithms to efficiently compute this derivative based on the adjoint method and reverse automatic differentiation (RAD) and RAD-based singular value derivative formula. Differentiation results for each method proposed were compared with FD results for one square and one tall rectangular matrix example and matched with the FD results to about 5–7 digits. Finally, we demonstrate the scalability of the proposed method by calculating the derivatives of singular values with respect to the snapshot matrix derived from the POD of a large dataset for a laminar–turbulent transitional flow over a flat plate, sourced from the John Hopkins turbulence database (JHTDB).
Journal
IF:
8.9
Papers:
1.3W
Citations:
6.6W
Organization
No organization information available
Cited Papers
Efficient Data-Driven Off-Design Constraint Modeling for Practical Aerodynamic Shape Optimization.png)
AIAA Journal
AIAA JournalIF0
PREAI
ShareSaveSensitivity-Based Geometric Parametrization and Automatic Scaling for Aerodynamic Shape OptimizationAIAA Journal
AIAA JournalIF0
PREAI
ShareSaveShareSaveShareSaveFrom Streaks to Spots and on to Turbulence: Exploring the Dynamics of Boundary Layer Transition
PREAI
ShareSaveShareSave
View more