Machine Learning Force Fields

forcefield

Overview

Machine Learning Force Fields (MLFFs) offer a novel approach for predicting the energies of arbitrary systems. While they have a slower runtime performance than classical force fields, they hold the potential to provide near-quantum accuracy at a fraction of the cost of DFT. MLFFs generally provide smooth potential energy functions that directly mimic DFT, depending only on the cartesian coordinates of the system. Some approaches also use the net charge and spin as input, allowing them to be applied to ionic and open-shell systems.

Current Limitations

Limitations for Molecular Dynamics Calculations:

Generally requires much smaller timesteps
No concepts of bonds/bond potentials means unintended "reactions" can occur, particularly in systems that begin with steric clashes
Currently no energy groups, so some analysis/property calculation doesn't work

Limitations for Quantum Mechanics Calculations:

Only first derivates are currently available analytically (geometry optimizations), second derivatives are computed by finite difference
Excited states are not available
Wavefunction properties are not available
- No coupling integrals, NMR, higher spin states (T1), open shell*, ESP charges

Available Models

MPNICE

Model Name	Description	Element Support	Levels of Theory	Application Area	Limitations
Organic_MPNICE	Trained to organic molecules	H C N O S F P Cl Br B (Supported but untested: Li Na Mg Si K Ca Sn I)	wB97X-D3BJ/def2-TZVPD	Organic molecules and liquids	Limited element coverage
Organic_MPNICE_TB	Delta learned model utilizing GFN2-xTB	H C N O S F P Cl Br B (Supported but untested: Li Na Mg Si K Ca I)	wB97X-D3BJ/def2-TZVPD	Finite system applications of organic molecules	Limited to finite systems
Organic_Crystals_MPNICE	Multi-task model trained to finite organic systems and organic crystals	H C N O S F P Cl Br	PBE-D3 (PW basis), wB97X-D3BJ/def2-TZVPD	Organic crystal structure optimization and ranking
Organic_Crystals_MPNICE_v1.1	Multi-task model trained to finite organic systems and organic crystals	H C N O S F Na Mg P Cl K Ca Zn Br I	PBE-D3 (PW basis), wB97X-D3BJ/def2-TZVPD	Crystal structure optimization and ranking, extended to hydrates and salts
Inorganic_MPNICE	Single task model trained to AIMD data from OMAT24 and fine tuned to MPTrj with GFN1-xTB charges	Up to Z=94, excluding Po, At, Rn, Fr, Ra	PBE-D3/GFN1-xTB charges	Bulk Inorganic Materials	Untested on surfaces, defects, organometallics. No magnetic properties.
Hybrid_MPNICE	Model with a single output head trained to MPtrj energies/forces/stress/GFN1-xTB charges), and the forces and dipoles for finite organic systems	Up to Z=94, excluding Po, At, Rn, Fr, Ra	wB97X-D3BJ/def2-TZVPD, PBE-D3/GFN1-xTB charges	Bulk inorganic materials, organic molecules	Interfaces between materials and organic molecules/liquids must be benchmarked versus a reference. Liquid densities for organics can have large error in density in NPT dynamics. Note that there is NO organometallic data in the training set
Hybrid_MPNICE_O	Multi task model trained to MPtrj and finite organic systems. This is the wB97X output head, which is slightly more accurate for organic tasks. Does not exactly reproduce wB97X total energies (uses PBE atomic energies in order to support full elemental coverage)	Up to Z=94, excluding Po, At, Rn, Fr, Ra	wB97X-D3BJ/def2-TZVPD, PBE-D3	Bulk inorganic materials, organic molecules and liquids	Interfaces between materials and organic molecules/liquids must be benchmarked versus a reference. Note that there is NO organometallic data in the training set
Hybrid_MPNICE_I	Multi task model trained to MPtrj and finite organic systems. This is the PBE-D3 output head, which is slightly more accurate for inorganic tasks such as mechanical properties of bulk inorganic crystals.	Up to Z=94, excluding Po, At, Rn, Fr, Ra	wB97X-D3BJ/def2-TZVPD, PBE-D3	Bulk inorganic materials, organic molecules and liquids	Interfaces between materials and organic molecules/liquids must be benchmarked versus a reference. Note that there is NO organometallic data in the training set

MPNICE does not have awareness of spin states above multiplicity 2, and the training data for the pretrained MPNICE models consisted of only singlet states.

UMA

The UMA models are developed by Meta and use a multi-task architecture with separate task heads for different datasets and levels of theory. UMA models support open-shell systems (multiplicity > 1) but do not have long-range interactions and are ~10x more expensive than MPNICE.

Model Name	Description	Element Support	Levels of Theory	Application Area
UMA_sm_omol	Trained on OMol25 (Open Molecules 2025)	Up to Z=82 (H–Pb)	wB97M-V/def2-TZVPD	Organic molecules, metal complexes, and electrolytes
UMA_sm_omc	Trained on OMC25 (Open Molecular Crystals 2025)	Up to Z=94, excluding Tb	PBE+D3 (PW)	Organic crystal structure optimization and ranking
UMA_sm_oc20	Trained on OC20 (Open Catalyst 2020)	Up to Z=94	RPBE (PW)	Catalysis and surface chemistry
UMA_sm_odac	Trained on ODAC23 (Open DAC 2023)	Up to Z=94	PBE-D3 (PW)	Adsorption on metal-organic frameworks
UMA_sm_omat	Trained on OMat24 (Open Materials 2024)	Up to Z=94, excluding Po, At, Rn, Fr, Ra	PBE/PBE+U (PW)	Bulk inorganic materials

The Schrödinger Suite includes the UMA small (sm) model, version 1.1. UMA can be run with the MLFF_inference.py script, Jaguar, and Desmond. To specify the model in MLFF_inference.py or Desmond, set the model type to UMA_sm_<task>. In Jaguar, all MLFF model names are specified in lowercase (for example, Organic_MPNICE becomes organic_mpnice); for UMA models, use mlff=uma_sm_<task>.

Custom

You can also supply your own trained MLFF model, generated by the MLFF Fine-tuning Panel. In a panel that provides a model menu, choose Custom, then use the Directory field and Browse button to select the folder that contains the model files.

Output Properties

MLFF calculations write energy and stress structure properties, per-atom forces, and, during molecular dynamics, per-atom velocities.

Property	Units
Energies (`r_sani_*energy`)	Hartree
Forces (`r_sani_fx`, `r_sani_fy`, `r_sani_fz`)	Hartree/Å
Stresses (`r_sani_stress_*`)	Hartree/Å³
Velocities (`r_ase_velocity_x`, `r_ase_velocity_y`, `r_ase_velocity_z`)	ASE internal units

To convert a velocity from ASE internal units to Å/fs, multiply it by approximately 0.09823.

Using MLFFs in place of Quantum Mechanics Calculations

An MLFF potential can be used for single point energy evaluations, geometry optimizations (constrained, TS optimizations), frequency calculations (numerical only currently), and scans. GPU acceleration is available for MPNICE and UMA models by setting gpu=1 in the Jaguar input file.

MLFF potentials can be selected from the Theory text box in Jaguar panels. You can select MLFF after clicking on the filter button to list all available models.

To run Jaguar with an MLFF potential from the command line, the following keywords can be added to Jaguar input file:

&gen
mlff=organic_mpnice_tb
mlff_solv=1 ! Turns on implicit solvent, only available *_tb models


mlff_solvent_model = ALPB ! solvent model to use, only applies if `mlff_solv=1`
&

See External Methods Keywords in the Jaguar Input File (Machine learning methods and xTB) for all keyword values.

To run batches of MLFF single point energy and geometry optimization calculations without launching a separate Jaguar job per structure, use the mlff.py: Batch MLFF Calculations script.

MLFFs in the Material Science Suite

MLFFs are available in various panels in the MS Suite:

MLFF Calculations Panel