Gibbs-Duhem-Informed Neural Networks for Binary Activity Coefficient Prediction

Gibbs–Duhem-informed neural networks provide a flexible hybrid approach to predicting binary activity coefficients with both high accuracy and thermodynamic consistency.

physics.chem-ph cs.LG Chemistry Computer Science Physics

Source

Paper

References (45)

01ML-SAFT: A machine learning framework for PCP-SAFT parameter predictionKobi C. Felton, Lukas Raßpe-Lange, J. Rittig et al.2024 · Chemical Engineering Journal · 31 citations

02HybridGamma: A thermodynamically consistent framework for hybrid modelling of activity coefficientsU. Di Caprio, J. Degrève, Peter Hellinckx et al.2023 · Chemical Engineering Journal · 11 citations

03Development of thermodynamically consistent machine-learning equations of state: Application to the Mie fluid.G. Chaparro, E. A. Müller2023 · Journal of Chemical Physics · 25 citations

04Leveraging uncertainty estimates and derivative information in Gaussian process regression for efficient collection and use of molecular simulation data.Jacob I Monroe, W. P. Krekelberg, Austin McDannald et al.2023 · Journal of Chemical Physics · 3 citations

05Predicting Temperature‐Dependent Activity Coefficients at Infinite Dilution Using Tensor CompletionJ. Damay, G. Ryzhakov, F. Jirasek et al.2023 · Chemie Ingenieur Technik · 13 citations

06Combining Machine Learning with Physical Knowledge in Thermodynamic Modeling of Fluid Mixtures.F. Jirasek, H. Hasse2023 · Annual Review of Chemical and Biomolecular Engineering · 64 citations

07Machine Learning for Predicting the Viscosity of Binary Liquid MixturesCamille L. Bilodeau, A. Kazakov, Sukrit Mukhopadhyay et al.2023 · Social Science Research Network · 64 citations

08Physics‐informed Transfer Learning for Out‐of‐sample Vapor Pressure PredictionsJoshua L. Lansford, K. Jensen, Brian Barnes2023 · Propellants, explosives, pyrotechnics · 19 citations

09SPT-NRTL: A physics-guided machine learning model to predict thermodynamically consistent activity coefficientsBenedikt Winter, Clemens Winter, Timm Esper et al.2022 · Fluid Phase Equilibria · 70 citations In Library

10Thermodynamically Consistent Vapor-Liquid Equilibrium Modelling with Artificial Neural NetworksAndrés Carranza Abaíd, H. Svendsen, J. Jakobsen2022 · Social Science Research Network · 15 citations

11Graph neural networks for materials science and chemistryPatrick Reiser, Marlen Neubert, Andr'e Eberhard et al.2022 · Communications Materials · 805 citations In Library

12Graph Neural Networks for Temperature-Dependent Activity Coefficient Prediction of Solutes in Ionic LiquidsJ. Rittig, K. B. Hicham, Artur M. Schweidtmann et al.2022 · Computers and Chemical Engineering · 72 citations In Library

Scroll for more · 33 remaining

Similar papers

Thermodynamics-Consistent Graph Neural NetworksJan G. Rittig, Alexander Mitsos2024 · arXiv (Cornell University) · 29 citations In Library

Gibbs-Helmholtz Graph Neural Network: capturing the temperature dependency of activity coefficients at infinite dilutionE. Medina, S. Linke, M. Stoll et al.2022 · Digital Discovery · 29 citations In Library

HANNA: Hard-constraint Neural Network for Consistent Activity Coefficient PredictionThomas Specht, Mayank Nagda, Sophie Fellenz et al.2024 · Chemical Science · 36 citations In Library

Machine Learning in Thermodynamics: Prediction of Activity Coefficients by Matrix CompletionF. Jirasek, Rodrigo Alves, J. Damay et al.2020 · Journal of Physical Chemistry Letters · 87 citations In Library

Graph Neural Networks for Temperature-Dependent Activity Coefficient Prediction of Solutes in Ionic LiquidsJ. Rittig, K. B. Hicham, Artur M. Schweidtmann et al.2022 · Computers and Chemical Engineering · 72 citations In Library

Graph neural networks informed locally by thermodynamicsAlicia Tierz, Ic´ıar Alfaro, David Gonz'alez et al.2025 · Engineering applications of artificial intelligence · 18 citations In Library

Making thermodynamic models of mixtures predictive by machine learning: matrix completion of pair interactionsF. Jirasek, Robert Bamler, Sophie Fellenz et al.2022 · Chemical Science · 27 citations In Library

Graph Neural Network-State Predictive Information Bottleneck (GNN-SPIB) approach for learning molecular thermodynamics and kineticsZiyue Zou, Dedi Wang, P. Tiwary2024 · Digital Discovery · 10 citations In Library

CoeffNet: predicting activation barriers through a chemically-interpretable, equivariant and physically constrained graph neural networkSudarshan Vijay, Maxwell C. Venetos, E. Spotte-Smith et al.2024 · Chemical Science · 5 citations In Library

Training Dynamical Binary Neural Networks with Equilibrium PropagationJérémie Laydevant, M. Ernoult, D. Querlioz et al.2021 · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · 22 citations In Library

Patents near this paper

Predicting Binary Outcomes Of An ActivityUS 11,120,103 · SALESFORCE.COM, INC. · 2021 APPROACHING HOMEOSTASIS IN A BINARY NEURAL NETWORKUS 11,138,493 · INTERNATIONAL BUSINESS MACHINES CORPORATION · 2021 SPIKING NEURAL NETWORK FOR PROBABILISTIC COMPUTATIONUS 11,449,735 · HRL LABORATORIES, LLC · 2022 NEURAL NETWORK ARCHITECTURE FOR GAUSSIAN COMPONENTS OF A MIXTURE DENSITY FUNCTIONUS 5,790,758 · 1998 NEURAL NETWORK ARCHITECTURE FOR NON-GAUSSIAN COMPONENTS OF A MIXTURE DENSITY FUNCTIONUS 5,712,959 · NAVY, UNITED STATES OF AMERICA, THE, AS REPRESENTED BY THE SECRETARY THE · 1998

Source of record

arXiv