Enum: MachineLearnedPotentialName

Machine learned potentials.

URI: https://CCPBioSim.ac.uk/biosim-schema/MachineLearnedPotentialName

Permissible Values

Value

Meaning

Description

MACE

None

Utilizes high-body-order expansions and provides exceptional force accuracy, …

ANI

None

Neural network potentials specifically parameterized for organic and small-mo…

NequIP

None

E(3)-equivariant network that models local atomic environments as tensors

UMA

None

Large-scale framework spanning half a billion atomic structures, and generali…

AceFF

None

Specifically optimized for small molecule drug discovery

Slots

Name

Description

machine_learned_potential_name

ML force field used to describe molecules

Identifier and Mapping Information

Schema Source

  • from schema: https://CCPBioSim.ac.uk/biosim-schema/

LinkML Source

name: MachineLearnedPotentialName
description: Machine learned potentials.
from_schema: https://CCPBioSim.ac.uk/biosim-schema/
rank: 1000
permissible_values:
  MACE:
    text: MACE
    description: Utilizes high-body-order expansions and provides exceptional force
      accuracy, excelling at capturing anharmonic dynamics in organic and macromolecular
      systems.
    aliases:
    - MACE
    - Multi-Atomic Cluster Expansion
  ANI:
    text: ANI
    description: Neural network potentials specifically parameterized for organic
      and small-molecule drug-like spaces (covering H, C, N, O).
    aliases:
    - ANI
    - Accurate Neural net Interaction
  NequIP:
    text: NequIP
    description: E(3)-equivariant network that models local atomic environments as
      tensors.
    aliases:
    - NequIP
    - Neural Equivariant Interatomic Potential
  UMA:
    text: UMA
    description: Large-scale framework spanning half a billion atomic structures,
      and generalizes across different chemical environments.
    aliases:
    - UMA
    - Universal Model for Atoms
  AceFF:
    text: AceFF
    description: Specifically optimized for small molecule drug discovery
    aliases:
    - AceFF