New N-GENNs framework learns complex thermodynamic dynamics from data

Researchers have introduced Nonlinear GENERIC-Embedded Neural Networks (N-GENNs), a novel deep learning framework designed to uncover the evolution equations for systems governed by the nonlinear GENERIC formalism. This approach effectively models systems with coupled conservative and dissipative dynamics by incorporating generalized gradient flows through convex dissipation potentials. The framework ensures strict adherence to thermodynamic laws by construction, enabling the identification of thermodynamically consistent dynamics, even with non-quadratic dissipation potentials. N-GENNs have been successfully validated on examples including a harmonic oscillator with a heat bath, a chemical motor, and a viscoplastic model, demonstrating their capability to accurately infer complex thermodynamic models from data. AI

IMPACT This framework could enable more accurate modeling of complex physical systems, potentially accelerating scientific discovery in fields relying on thermodynamics.

RANK_REASON Academic paper introducing a new computational method. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• General Equation for Non-Equilibrium Reversible-Irreversible Coupling
• N-GENNs
• Nonlinear GENERIC-Embedded Neural Networks
• Vojtěch Votruba

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →