Update README.md with new RSS functionalities and clean up #202
Conversation
README.md
Outdated
| We currently have two different types of workflows available. | ||
| * Workflow to use random-structure searches for the systematic construction of interatomic potentials | ||
| * Workflow to train accurate interatomic potentials for harmonic phonon properties |
There was a problem hiding this comment.
@vlderinger Fine like this?
There was a problem hiding this comment.
Sounds great! We could say "iterative random structure searching (RSS)" to make it clearer that the process runs over many loops – this is where the workflows are particularly useful. @YuanbinLiu are you happy with the above wording too?
I would also link here to the key publications for the ideas that are implemented in the respective workflows – https://doi.org/10.1103/PhysRevLett.120.156001 and https://doi.org/10.1038/s41524-019-0236-6 for GAP-RSS, and https://doi.org/10.1063/5.0013826 for phonons.
There was a problem hiding this comment.
Yes, sounds great to me!
There was a problem hiding this comment.
@QuantumChemist Could you add the references mentioned above to the readme as well? :). Thanks in advance.
|
Would you mind if I smuggled in here the one commit I want to make to add to the documentation that the jobnames within autoplex will be fixed? |
|
@QuantumChemist i would not. Go ahead |
Okidki 😃 |
|
@QuantumChemist, Could you (next week) please add a disclaimer to the README.md stating that we currently expect familiarity with jobflow, atomate2, and jobflow-remote/fireworks and that people should, for now, refer to the respective documentation for further information? We could also add this to the documentation |
I will add all the commits then to this PR directly! |
README.md
Outdated
| After a few data preprocessing steps (`data preprocessing`) to filter out data with too strong force values, the MLIP fit (`machine learning fit`) is run and the resulting potential is used for the benchmark against DFT data (`benchmark`). | ||
| Finally, the result metrics are collected in form of output plots and files (`write benchmark metrics`). | ||
| We currently have two different types of workflows available. | ||
| * Workflow to use random-structure searches for the systematic construction of interatomic potentials ([[1]](https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.120.156001), [[2]](https://www.nature.com/articles/s41524-019-0236-6)) |
There was a problem hiding this comment.
I think I would put in a typical journal-like citation here and add the link in the background. Also, we.should emphasize that these are the ideas that we are automating. Otherwise, it might look like the automation already exists.
|
@JaGeo , could you let me know if you like the changes I made or if there is something else to add and change? Linting is failing because of "Mater" in the npj ref |
Co-authored-by: J. George <JaGeo@users.noreply.github.com>
|
Nothing worked, except for shortening it further to "Mat." 🙈 |
|
@QuantumChemist, you can add exceptions to codespell in the pre-commit configuration file. Then, it works :). I did that and added a few tiny changes to the file as well. In any case: thank you! @vlderinger I will merge this for now and we can still implement further suggestions in a new PR. |
Hi @vlderinger @YuanbinLiu @QuantumChemist ,
I started with the pull request. I am happy to work together on improving the README.md to cover the RSS functionality in a better way.