---

---

ArcaNN

ArcaNN proposes an automated enhanced sampling generation of training sets for chemically reactive machine learning interatomic potentials. In its current version, it aims to simplify and automate the iterative training process of a DeePMD-kit neural network potential for a user-chosen system. The core concepts of this training procedure could be extended to other network architectures.

Key Advantages

• Modularity: The code is designed with modularity in mind, allowing users to finely tune the training process to fit their specific system and workflow.
• Traceability: Every parameter set during the procedure is recorded, ensuring great traceability.

Iterative Training Process

During the iterative training process, you will:

1. Train neural network potentials.

2. Use them as reactive force fields in molecular dynamics simulations to explore the phase space.

3. Select and label configurations based on a query by committee approach.

4. Train a new generation of neural network potentials again with the improved training set.

This workflow, often referred to as active or concurrent learning, is inspired by DP-GEN. We adopt their naming scheme for the steps in the iterative procedure. Each iteration, or cycle, consists of the following steps:

• Training
• Exploration
• Labeling
• (Optional) Testing

Ensure you understand the meaning of each step before using the code.

GitHub Repository Structure

You will find in our GitHub repository everything you need to set up the ArcaNN software, as well as example files that you can use as an example. The repository contains several folders:

• The tools/ folder contains helper scripts and files.
• The arcann_training/ folder contains the ArcaNN Training code. We strongly advise against modifying its contents. See ArcaNN Installation Guide for the installation.
• The examples/ folder contains template files to set up the iterative training procedure for your system. Within this folder you can find :
• The inputs/ folder with five JSON files, one per step. These files contain all the keywords used to control each step of an iteration (namely initialization, exploration, labeling, training and optionally test), including their type and the default values taken by the code if a keyword isn't provided by the user. If the default is a list containing a single value it means that this value will be repeated and used for every system (see the corresponding section fir each step). Note : For the exploration step some keywords have two default values, the first one will be used if the exploration is conducted with classical nuclei MD (i.e. LAMMPS) and the second one will be used with quantum nuclei MD (i.e. i-PI).
• The user_files/ folder with:
  • A machine.json template file where all the information about your cluster should be provided (see HPC Configuration).
  • A input folder for each step, where skeleton files are provided as templates for writing your own inputs for the respective external programs. For example, in the exploration_lammps/, labeling_cp2k/ and training_deepmd/, you can find the necessecary files to perform the exploration with LAMMPS, the labeling with CP2K and the training with DeePMD-kit (see Exploration, Labeling and Training for a detailed description of the tunneable keywords).
  • A job folder for each step, where skeleton submission files are provided as template that ArcaNN use to launch the different phases in each step when they require a HPC machine. For example, in job_exploration_lammps_slurm/, job_labeling_CP2K_slurm, job_training_deepmd_slurm and the optional step job_test_deepmd_slurm, you can find basic Slurm submission files.

You must adapt these files to ensure they work on your machine (see Usage), but be careful not to modify the replaceable keywords (every word starting with _R_ and ending with _) that Arcann will replace with user-defined or auto-generated values (e.g., the wall time for labeling calculations, the cluster partition to be used, etc.).