Geometry-agnostic model reduction with GNN-generated reduced POD bases and boosted PGD enrichment for (non)linear structural elastodynamics
Bibliographic record
Abstract
This contribution proposes a new and significantly enhanced extension of a recently-introduced hybrid Graph Neural Network (GNN)-based reduced-order modeling approach for the numerical solution of time-dependent partial differential equations on non-parametric finite element meshes. Building upon previous proof-of-concept work, this more generalized framework presents a number of key novelties: tight integration of graph-based learning with physical information via direct imposition of finite element operators as node and edge level features; introduction of a Grassmannian subspace distance measure as a dedicated training objective; incorporation of a Gated Recurrent Unit (GRU) for a more efficient and lightweight architecture; hybridization with other Galerkin-based reduced-order methods such as the Proper Orthogonal Decomposition (POD); and a first treatment of nonlinear problems. A novel, on-the-fly enrichment mechanism, modified from a classical Proper General Decomposition (PGD) and dubbed ”Boosted PGD”, is additionally introduced to improve prediction accuracy at low computational cost via additional greedy corrective modes. The efficacy of the overall methodology is assessed on two challenging datasets featuring significant geometric and topological variations that include highly heterogeneous spatial discretizations. A variety of performance studies demonstrate very competitive accuracy and computational cost in simulating highly-dynamic behavior when compared to conventional full-order finite element models, including a remarkable capacity to generalize to configurations well outside of the topological scope of the original training and validation sets. Results imply that solvers constructed from such an approach may enable more scalable and robust mechanical simulations for complex, real-world engineering applications related to iterative design.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".