marsalekgroup / aml

A... M... L...

AML

Short Description

This is a Python package to automatically build the reference set for the training of Neural Network Potentials (NNPs), and eventually other machine-learned potentials, in an automated, data-driven fashion. For that purpose, a large set of reference configurations sampled in a physically meaningful way (typically with molecular dynamics) is filtered and the most important points for the representation of the Potential Energy Surface (PES) are identified. This is done by using a set of NNPs, called a committee, for error estimates of individual configurations. By iteratively adding the points with the largest error in the energy/force prediction, the reference set is progressively extended and optimized.

Keywords:

• Active learning
• Query by committee
• Ensemble averaging
• Committee machines
• Neural Network Potentials

More information can be found in the following references:

• C. Schran, F. L. Thiemann, P. Rowe, E. A. Müller, O. Marsalek, A. Michaelides,
  "Machine learning potentials for complex aqueous systems made simple",
  PNAS 118, e2110077118 (2021), 10.1073/pnas.2110077118
• C. Schran, K. Brezina, O. Marsalek,
  "Committee neural network potentials control generalization errors and enable active learning",
  J. Chem. Phys. 153, 104105 (2020), 10.1063/5.0016004

Installation

For now, just clone the repository and source the env.sh file.

Dependencies:

• NumPy
• SciPy
• Matplotlib
• MDTraj (for scoring)
• cp2k-input-tools (for CP2K calculations)

aml's Issues

Exception: No weights file found for element: O

Dear AML Team,

I just want to use the aml to generate a NNP potential of water. When I used the script of "create-model.py" prepared in the examples, I encountered an issue like below:

Traceback (most recent call last):
File "/lustre1/u/u3004964/using_n2p2/aml-committee/water-cheng/create-model.py", line 35, in
n2p2.train(structures, n_epoch=2)
File "/lustre1/u/u3004964/software/aml/aml/mlp.py", line 206, in train
self._read_parameters(directories)
File "/lustre1/u/u3004964/software/aml/aml/mlp.py", line 160, in _read_parameters
model_parameters.append(self._read_output_train(directory))
File "/lustre1/u/u3004964/software/aml/aml/mlp.py", line 449, in _read_output_train
raise Exception(f'No weights file found for element: {element:s}')
Exception: No weights file found for element: O

I carefully prepared the input files (including input.data, input.nn and scaling,data) for training the NNP by using n2p2, and I also tested and found that it is correct if I directly run 'nnp-train' in the path of 'final-training/train-000'. Could you tell me what should I do for this issue? Many thanks.

I am also confused about the aml function, does aml can be sued to train the NNP potential? or we need to use the n2p2 to train and then use the aml to generate the C-NNP?

Cheers,
zezhu

Continuing Incomplete Jobs

Hello,

I have been trying to run QbC jobs with AML, but have found that long jobs stop halfway through with no errors or logs explaining why.

I have yet to find the cause of this issue.

I was wondering if there is some way that I can continue an incomplete QbC job.

For example, the most recent job stopped at iteration 8 of 10. Is there some way that I can rerun the QbC for just the last 3 iterations?

Thanks,
Oliver

The energy and force values won't update

Hello,
I was playing around with the example to run a Qbc for a system of mine. My energy and force files are in CP2K format, and when I run the run-Qbc.py it generates the RuNNer format file properly. However, when I look at the "eval_train" and "predict" folders, I can see that the values of energy and forces in the "input.data" file are zero. Would you please help me find out what could go wrong?

Thank you

ModuleNotFoundError: No module named 'mdtraj'

Hello, I am trying to use the "score-all.py" to validate my trajectories but this error pops up
" ModuleNotFoundError: No module named 'mdtraj' "
I tried to replace the line " import mdtraj as mdt " with " from aml.io import mdtraj as mdt ", but then I get another error
"AttributeError: module 'aml.io.mdtraj' has no attribute 'load_frame' "
I was wondering if you could help me fix this issue.

Cheers
Mozhdeh