mlsanityData to 7 Oct 2026

Thermodynamics-Consistent Graph Neural Networks

arXiv 2407.18372 ↗First posted 8 Jul 2024cond-mat.dis-nn cs.LG physics.chem-phcited by 12

Not on the map. Why?

Abstract (arXiv)

We propose excess Gibbs free energy graph neural networks (GE-GNNs) for predicting composition-dependent activity coefficients of binary mixtures. The GE-GNN architecture ensures thermodynamic consistency by predicting the molar excess Gibbs free energy and using thermodynamic relations to obtain activity coefficients. As these are differential, automatic differentiation is applied to learn the activity coefficients in an end-to-end manner. Since the architecture is based on fundamental thermodynamics, we do not require additional loss terms to learn thermodynamic consistency. As the output is a fundamental property, we neither impose thermodynamic modeling limitations and assumptions. We demonstrate high accuracy and thermodynamic consistency of the activity coefficient predictions.

Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Builds on 11 of its 38 references that are in the map. First posted on arXiv 8 Jul 2024.

Builds on

11 of 38 references.

References · 11

  1. 2017Neural Message Passing for Quantum Chemistrycited by 3,460
  2. 2022Graph neural networks for materials science and chemistrycited by 181
  3. 2022A smile is all you need: Predicting limiting activity coefficients from SMILES with natural language processingcited by 19
  4. 2023Gibbs-Duhem-Informed Neural Networks for Binary Activity Coefficient Predictioncited by 17
  5. 2022Graph Neural Networks for Temperature-Dependent Activity Coefficient Prediction of Solutes in Ionic Liquidscited by 14
  6. 2022Graph neural networks for the prediction of molecular structure-property relationshipscited by 12
  7. 2022Gibbs-Helmholtz Graph Neural Network: capturing the temperature dependency of activity coefficients at infinite dilutioncited by 12
  8. 2022SPT-NRTL: A physics-guided machine learning model to predict thermodynamically consistent activity coefficientscited by 11
  9. 2022Physical Pooling Functions in Graph Neural Networks for Molecular Property Predictioncited by 7
  10. 2024Balancing Molecular Information and Empirical Data in the Prediction of Physico-Chemical Propertiescited by 3
  11. 2018Learning representations of molecules and materials with atomistic neural networkscited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 4; 2025: 5; 2026: 3.

12 papers.

  1. 2024HANNA: Hard-constraint Neural Network for Consistent Activity Coefficient Predictioncited by 17
  2. 2024GraphXForm: Graph transformer for computer-aided molecular designcited by 6
  3. 2025Molecular Machine Learning in Chemical Process Designcited by 5
  4. 2024Hierarchical Matrix Completion for the Prediction of Properties of Binary Mixturescited by 4
  5. 2024Predicting the Temperature-Dependent CMC of Surfactant Mixtures with Graph Neural Networkscited by 3
  6. 2025Thermodynamically consistent machine learning model for excess Gibbs energycited by 3
  7. 2025Graph Neural Networks embedded into Margules model for vapor-liquid equilibria predictioncited by 2
  8. 2025Federated Learning from Molecules to Processes: A Perspectivecited by 2
  9. 2026Differentiable Thermodynamic Phase-Equilibria for Machine Learningcited by 1
  10. 2025DeepEOSNet: Capturing the dependency on thermodynamic state in property prediction taskscited by 0
  11. 2026Clapeyron Neural Networks for Single-Species Vapor-Liquid Equilibriacited by 0
  12. 2026A Systematic Evaluation of Molecular Mixture Behavior Predictioncited by 0