A reduced-order modeling framework that learns latent dynamics with enforced energy conservation and entropy production, using parametric GENERIC-informed networks, and reports up to 3,528x speedup on Burgers and Vlasov-Poisson benchmarks.
Review of digital twin about concepts, technologies, and industrial applications,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Thermodynamically Consistent Latent Dynamics Identification for Parametric Systems
A reduced-order modeling framework that learns latent dynamics with enforced energy conservation and entropy production, using parametric GENERIC-informed networks, and reports up to 3,528x speedup on Burgers and Vlasov-Poisson benchmarks.