ReactionAtlas is an iterative ML framework that proposes candidate reactions from seed molecules, filters them with an ML force field for valid transition states, and grows a network of ~47,000 reactions among ~12,000 compounds up to C4 in pre-biotic chemistry.
hub Mixed citations
Generalized neural-network representation of high-dimensional potential-energy surfaces
Mixed citation behavior. Most common role is background (50%).
19
Pith papers citing it
4,276
external citations · Crossref
Background
50% of classified citations
hub tools
citation-role summary
background 3
method 2
baseline 1
citation-polarity summary
representative citing papers
LADeQ is an LLM-driven workflow that autonomously discovers and implements approximation algorithms for CCSD and CISD calculations, delivering speedups while respecting user-specified error tolerances.
Stable size extrapolation in local score models requires the receptive field to cover the quasi-locality range of the Gaussian-smoothed score, formalized via a size-uniform comparison theorem and validated on the new FDLF benchmark.
A new polarizable QM/MM method for periodic systems uses SCME for water with multipoles up to hexadecapole and anisotropic polarizabilities, achieving full QM accuracy via careful near/far-field expansions and damping.
Kernels from pretrained MLIP latent spaces outperform standard acquisition methods in active learning for reactive chemistry, reducing required labels by 38% for energy error and 28% for force error.
Machine-learned many-body potentials from Poisson-Boltzmann calculations on clusters up to 48 colloids show that higher-order interactions reduce cohesion and eliminate broad gas-liquid phase separation, consistent with primitive model pair and triplet potentials.
Force-aware Neural Tangent Kernels combined with chunked acquisition provide scalable and distribution-robust active learning for MLIPs, outperforming baselines on OC20 and remaining competitive on other benchmarks.
DenSNet learns the Hohenberg-Kohn map to electron density with equivariant networks and delta-learning, then maps density to energy, producing stable MD trajectories whose infrared spectra match experiment and DFT on ethanol, ethanethiol, resorcinol, and polythiophene oligomers.
A 1.62-trillion-atom molecular dynamics simulation achieves ab initio accuracy with 100x speedup over prior machine learning force fields and 86.9% weak scaling to 45,000 GPGPUs.
Bayesian E(3)-equivariant MLPs with joint energy-force NLL loss achieve competitive accuracy while enabling uncertainty-guided active learning, OOD detection, and calibration.
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.
DeltaDiff is a physics-guided inference method that predicts mutant protein structures from a baseline diffusion model without retraining, tested on three systems with nonlocal changes.
Systematic benchmarking finds Grønbech-Jensen-Farago Langevin thermostat most consistent for temperature and energy sampling in binary LJ glass simulations, at roughly double the cost and with friction-dependent diffusion.
Machine-learned force fields trained on coupled-cluster potential energy surfaces produce phonon dispersions and vibrational densities of states for solids that agree better with experiment than DFT-based models.
Neural network approximates potential from Hamiltonian trajectories then equation discovery extracts algebraic expression matching ground truth on oscillators, central force, and Coulomb problems.
ML-potential MD simulations of sodium disilicate, tetrasilicate and hexasilicate melts show sodium hopping via bimodal van Hove functions and strongest non-Gaussian parameter for oxygen atoms.
Bayesian optimization with Gaussian processes unifies minimization, single-point saddle searches, and double-ended path searches on potential energy surfaces through a shared six-step surrogate loop using derivative observations and inverse-distance kernels.
This perspective article develops a definition of foundational MLIPs and poses six open questions that the authors believe will define future research in machine-learned interatomic potentials.
citing papers explorer
-
ReactionAtlas: Ab origine exploration of chemical reaction networks with machine learning
ReactionAtlas is an iterative ML framework that proposes candidate reactions from seed molecules, filters them with an ML force field for valid transition states, and grows a network of ~47,000 reactions among ~12,000 compounds up to C4 in pre-biotic chemistry.
-
LLM-Guided Test-Time Discovery of Quantum-Chemical Approximation Algorithms
LADeQ is an LLM-driven workflow that autonomously discovers and implements approximation algorithms for CCSD and CISD calculations, delivering speedups while respecting user-specified error tolerances.
-
When Do Local Score Models Extrapolate Across Size? A Diagnostic Theory and Benchmark
Stable size extrapolation in local score models requires the receptive field to cover the quasi-locality range of the Gaussian-smoothed score, formalized via a size-uniform comparison theorem and validated on the new FDLF benchmark.
-
Polarizable Embedding QM/MM for Periodic Systems
A new polarizable QM/MM method for periodic systems uses SCME for water with multipoles up to hexadecapole and anisotropic polarizabilities, achieving full QM accuracy via careful near/far-field expansions and damping.
-
Pretrained Model Representations as Acquisition Signals for Active Learning of MLIPs
Kernels from pretrained MLIP latent spaces outperform standard acquisition methods in active learning for reactive chemistry, reducing required labels by 38% for energy error and 28% for force error.
-
Many-body attractions do not stabilize gas-liquid phase separation in aqueous dispersions of charged colloids within the Poisson-Boltzmann framework
Machine-learned many-body potentials from Poisson-Boltzmann calculations on clusters up to 48 colloids show that higher-order interactions reduce cohesion and eliminate broad gas-liquid phase separation, consistent with primitive model pair and triplet potentials.
-
Force-Aware Neural Tangent Kernels for Scalable and Robust Active Learning of MLIPs
Force-aware Neural Tangent Kernels combined with chunked acquisition provide scalable and distribution-robust active learning for MLIPs, outperforming baselines on OC20 and remaining competitive on other benchmarks.
-
Enhancing molecular dynamics with equivariant machine-learned densities
DenSNet learns the Hohenberg-Kohn map to electron density with equivariant networks and delta-learning, then maps density to energy, producing stable MD trajectories whose infrared spectra match experiment and DFT on ethanol, ethanethiol, resorcinol, and polythiophene oligomers.
-
Trillion-atom molecular dynamics simulations with ab initio accuracy
A 1.62-trillion-atom molecular dynamics simulation achieves ab initio accuracy with 100x speedup over prior machine learning force fields and 86.9% weak scaling to 45,000 GPGPUs.
-
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.
-
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.
-
DeltaDiff: Training-Free, Physics-Guided Machine Learning for Predicting Mutant Protein Structures
DeltaDiff is a physics-guided inference method that predicts mutant protein structures from a baseline diffusion model without retraining, tested on three systems with nonlocal changes.
-
Benchmarking thermostat algorithms in molecular dynamics simulations of a binary Lennard-Jones glass-former model
Systematic benchmarking finds Grønbech-Jensen-Farago Langevin thermostat most consistent for temperature and energy sampling in binary LJ glass simulations, at roughly double the cost and with friction-dependent diffusion.
-
Machine-Learned Force Fields for Lattice Dynamics at Coupled-Cluster Level Accuracy
Machine-learned force fields trained on coupled-cluster potential energy surfaces produce phonon dispersions and vibrational densities of states for solids that agree better with experiment than DFT-based models.
-
Learning and Interpreting Potentials for Classical Hamiltonian Systems
Neural network approximates potential from Hamiltonian trajectories then equation discovery extracts algebraic expression matching ground truth on oscillators, central force, and Coulomb problems.
-
Dynamic heterogeneity in sodium silicate melts via machine-learning potential
ML-potential MD simulations of sodium disilicate, tetrasilicate and hexasilicate melts show sodium hopping via bimodal van Hove functions and strongest non-Gaussian parameter for oxygen atoms.
-
A Tutorial Review of Bayesian Optimization with Gaussian Processes to Accelerate Stationary Point Searches
Bayesian optimization with Gaussian processes unifies minimization, single-point saddle searches, and double-ended path searches on potential energy surfaces through a shared six-step surrogate loop using derivative observations and inverse-distance kernels.
-
Six Open Questions in Machine-Learned Interatomic Potential Foundation Models
This perspective article develops a definition of foundational MLIPs and poses six open questions that the authors believe will define future research in machine-learned interatomic potentials.
- Economised path integrals