Introduces graph-to-image prediction of per-node dynamic stability landscapes in oscillator networks from topology, releases two 10k-graph datasets, and shows GNN-CNN models achieve good accuracy with cross-size generalization.
hub
Directional
17 Pith papers cite this work, alongside 352 external citations. Polarity classification is still indexing.
17
Pith papers citing it
352
external citations · Pith
abstract
Graph neural networks have recently achieved great successes in predicting quantum mechanical properties of molecules. These models represent a molecule as a graph using only the distance between atoms (nodes). They do not, however, consider the spatial direction from one atom to another, despite directional information playing a central role in empirical potentials for molecules, e.g. in angular potentials. To alleviate this limitation we propose directional message passing, in which we embed the messages passed between atoms instead of the atoms themselves. Each message is associated with a direction in coordinate space. These directional message embeddings are rotationally equivariant since the associated directions rotate with the molecule. We propose a message passing scheme analogous to belief propagation, which uses the directional information by transforming messages based on the angle between them. Additionally, we use spherical Bessel functions and spherical harmonics to construct theoretically well-founded, orthogonal representations that achieve better performance than the currently prevalent Gaussian radial basis representations while using fewer than 1/4 of the parameters. We leverage these innovations to construct the directional message passing neural network (DimeNet). DimeNet outperforms previous GNNs on average by 76% on MD17 and by 31% on QM9. Our implementation is available online.
hub tools
citation-role summary
background 2
method 1
citation-polarity summary
representative citing papers
JanusPipe introduces SymFold and WaveK to enable efficient 3D-parallel training for conservative MLIPs, reporting 1.51x and 1.45x average throughput gains over 1F1B and Hanayo baselines on 32 GPUs.
Chem-GMNet uses sphere-native embeddings, DualSKA attention, and SH-FFN layers to match or beat ChemBERTa-2 on MoleculeNet tasks with fewer parameters and sometimes no pretraining.
QT-Net predicts atomic electron populations and multipoles via a new SOAP-cluster held-out test, improving molecular property prediction and recovering QM9 dipole moments from per-atom outputs.
A composition-weighted symbolic regression framework learns analytical expressions and elemental weightings from composition to predict materials properties with accuracy competitive to black-box models while producing explicit, constraint-enforcing formulas.
Skala is a neural XC functional trained on wavefunction data that beats state-of-the-art hybrids on main-group chemistry benchmarks at semi-local computational cost.
pDOS-augmented GNN node features reduce Tc and ε∞ prediction errors by 22.9% and 27.9% versus elemental baselines, matching roughly a 1.7–1.8× increase in training data.
A lightweight Pairformer architecture learns accurate biomolecular committors from atom-level pairs, revealing chignolin TSE bifurcation and substituent-tuned host–guest binding pathways.
A dual-stream crystal graph network with angle-aware metal–ligand–metal exchange features lowers magnetic-moment prediction MAE from 2.54 to 2.02 μB and, after pretraining, improves FM/AFM classification accuracy from 73.6% to 75.7%.
LLM molecular design framework uses self-reflection on full physicochemical data from first-principles calculations to achieve low deviation on HOMO-LUMO gaps and generalize to other properties.
SurfDesign introduces surface-conditioned protein design via manifold modeling and equivariant message passing on surfaces integrated with pretrained language models, outperforming prior methods on binder and enzyme design benchmarks.
Bayesian E(3)-equivariant MLPs with joint energy-force NLL loss achieve competitive accuracy while enabling uncertainty-guided active learning, OOD detection, and calibration.
Geometric deep learning provides a unified mathematical framework based on grids, groups, graphs, geodesics, and gauges to explain and extend neural network architectures by incorporating physical regularities.
Transfer learning from QM9 with transformers and uncertainty quantification predicts Hansen solubility parameters, dielectric constants, and limited-data Gutmann numbers to enable green solvent screening.
Loss-guided adaptive scale refinement on NaCl aqueous system reduces overall force MAE from 399.65 to 381.23 by discovering intermediate scales from initial anchors.
OCVF adds a learned neural-network correction to a skeleton Hamiltonian so that the model matches experimental PDF constraints, yielding up to 95.8% better accuracy on BaTiO3 phase-transition temperatures.
citing papers explorer
-
Learning Dynamic Stability Landscapes in Synchronization Networks
Introduces graph-to-image prediction of per-node dynamic stability landscapes in oscillator networks from topology, releases two 10k-graph datasets, and shows GNN-CNN models achieve good accuracy with cross-size generalization.
-
JanusPipe: Efficient Pipeline Parallel Training for Machine Learning Interatomic Potentials
JanusPipe introduces SymFold and WaveK to enable efficient 3D-parallel training for conservative MLIPs, reporting 1.51x and 1.45x average throughput gains over 1F1B and Hanayo baselines on 32 GPUs.
-
Chem-GMNet: A Sphere-Native Geometric Transformer for Molecular Property Prediction
Chem-GMNet uses sphere-native embeddings, DualSKA attention, and SH-FFN layers to match or beat ChemBERTa-2 on MoleculeNet tasks with fewer parameters and sometimes no pretraining.
-
QT-Net: Rethinking Evaluation of AI Models in Atomic Chemical Space
QT-Net predicts atomic electron populations and multipoles via a new SOAP-cluster held-out test, improving molecular property prediction and recovering QM9 dipole moments from per-atom outputs.
-
Composition-Weighted Symbolic Regression for General-Purpose Property Prediction
A composition-weighted symbolic regression framework learns analytical expressions and elemental weightings from composition to predict materials properties with accuracy competitive to black-box models while producing explicit, constraint-enforcing formulas.
-
Accurate and scalable exchange-correlation with deep learning
Skala is a neural XC functional trained on wavefunction data that beats state-of-the-art hybrids on main-group chemistry benchmarks at semi-local computational cost.
-
Machine Learning Materials Properties by Encoding Orbital-Projected Density of States
pDOS-augmented GNN node features reduce Tc and ε∞ prediction errors by 22.9% and 27.9% versus elemental baselines, matching roughly a 1.7–1.8× increase in training data.
-
Navigating committor landscape of biomolecules with a general pairwise interaction model
A lightweight Pairformer architecture learns accurate biomolecular committors from atom-level pairs, revealing chignolin TSE bifurcation and substituent-tuned host–guest binding pathways.
-
mCGCNN: A Dual-Stream Crystal Graph Convolutional Neural Network for the Efficient Prediction of Magnetic Properties of Crystalline Materials
A dual-stream crystal graph network with angle-aware metal–ligand–metal exchange features lowers magnetic-moment prediction MAE from 2.54 to 2.02 μB and, after pretraining, improves FM/AFM classification accuracy from 73.6% to 75.7%.
-
Closing the Prior-Posterior Loop: Self-Reflective Molecular Design with Analysis-Driven LLM Iteration
LLM molecular design framework uses self-reflection on full physicochemical data from first-principles calculations to achieve low deviation on HOMO-LUMO gaps and generalize to other properties.
-
SurfDesign: Effective Protein Design on Molecular Surfaces
SurfDesign introduces surface-conditioned protein design via manifold modeling and equivariant message passing on surfaces integrated with pretrained language models, outperforming prior methods on binder and enzyme design benchmarks.
-
Bayesian E(3)-Equivariant Interatomic Potential with Iterative Restratification of Many-body Message Passing
Bayesian E(3)-equivariant MLPs with joint energy-force NLL loss achieve competitive accuracy while enabling uncertainty-guided active learning, OOD detection, and calibration.
-
Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
Geometric deep learning provides a unified mathematical framework based on grids, groups, graphs, geodesics, and gauges to explain and extend neural network architectures by incorporating physical regularities.
-
A green solvent screening tool for emerging materials via uncertainty aware, transformer enhanced transfer learning
Transfer learning from QM9 with transformers and uncertainty quantification predicts Hansen solubility parameters, dielectric constants, and limited-data Gutmann numbers to enable green solvent screening.
-
Loss-Guided Adaptive Scale Refinement for Molecular Force Prediction
Loss-guided adaptive scale refinement on NaCl aqueous system reduces overall force MAE from 399.65 to 381.23 by discovering intermediate scales from initial anchors.
-
Infusing Experimental Reality into Complex Many-Body Hamiltonians: The Observable-Constrained Variational Framework (OCVF)
OCVF adds a learned neural-network correction to a skeleton Hamiltonian so that the model matches experimental PDF constraints, yielding up to 95.8% better accuracy on BaTiO3 phase-transition temperatures.
- BiScale-GTR: Fragment-Aware Graph Transformers for Multi-Scale Molecular Representation Learning