Standard model outputs

In order for multiple simulation engines to be able use arbitrary metatomic models to compute atomic properties, we need all the models to specify the same metadata for a given output. If your model returns one of the outputs defined in this documentation, then it should follow the metadata structure described here.

If you need other outputs, you should use custom output with a name containing ::, such as my_code::my_output. For such custom outputs, you are free to use any relevant metadata structure, but if multiple people are using the same kind of outputs, they are encouraged to come together, define the metadata schema they need and add a new section to these pages.

Output variants

Models can define variants of any output, for example to provide the same output at different levels of theory in a single model. For more information on output variants, please refer to the corresponding documentation.

Physical quantities

The first set of standardized outputs for metatomic models are physical quantities, i.e. quantities with a well-defined physical meaning.

Energy
../_images/energy-output.png

The potential energy associated with a given system configuration. This can be used to run molecular simulations with on machine learning based interatomic potentials.

Energy
Energy ensemble
../_images/energy-ensemble-output.png

An ensemble of multiple potential energy predictions, generated when running multiple models simultaneously.

Energy ensemble
Energy uncertainty
../_images/energy-uncertainty-output.png

The uncertainty on the potential energies, useful to quantify the confidence of the model.

Energy uncertainty
Non-conservative forces
../_images/nc-forces-output.png

Forces directly predicted by the model, not derived from the potential energy.

Non-conservative forces
Non-conservative stress
../_images/nc-stress-output.png

Stress directly predicted by the model, not derived from the potential energy.

Non-conservative stress
Masses
../_images/masses-output.png

Atomic masses

Masses
Positions
../_images/positions-output.png

Atomic positions predicted by the model, to be used in ML-driven simulations.

Positions
Momenta
../_images/momenta-output.png

Atomic momenta, i.e. \(m \times \vec v\)

Momenta
Velocities
../_images/velocities-output.png

Atomic velocities, i.e. \(\vec p / m\)

Velocities

Machine learning quantities

The next set of standardized outputs in metatomic models are specific to machine learning and related tools.

Features
../_images/features-output.png

Features are numerical vectors representing a given structure or atomic environment in an abstract n-dimensional space.

Features