A transformer model with self-attention and auxiliary physics losses learns a direct non-iterative mapping from loads and fields to manufacturable optimized topologies.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CE 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Physics-Informed Transformer for Real-Time High-Fidelity Topology Optimization
A transformer model with self-attention and auxiliary physics losses learns a direct non-iterative mapping from loads and fields to manufacturable optimized topologies.