Project ID: plumID:20.004
Name: Data-driven collective variables for enhanced sampling
Archive: (browse)
Category: methods
Keywords: collective variables, machine learning, deep-lda
PLUMED version: 2.5-dev
Contributor: Luigi Bonati
Submitted on: 18 Feb 2020
Last revised: 06 Apr 2020
Publication: L. Bonati, V. Rizzi, M. Parrinello, Data-Driven Collective Variables for Enhanced Sampling. The Journal of Physical Chemistry Letters. 11, 2998–3004 (2020)

PLUMED input files

File Compatible with
ala2/3_enhanced_sampling/plumed.dat tested on v2.9 tested on master
aldol/3_enhanced_sampling/plumed.dat tested on v2.9 tested on master with LOAD

Last tested: 18 Jun 2024, 14:35:20

Project description and instructions
This egg contains the Pytorch interface and input files necessary to use a Deep-LDA collective variable, as well as additional code to reproduce the result reported in the paper. The Deep-LDA CV is constructed by optimizing a neural network using Fisher’s linear discriminant as the objective function to maximize the separation between the states. Note that the build fails because PLUMED needs to be manually compiled with LibTorch. We have released a Google Colab Notebook, available here with a tutorial about the Deep-LDA CV training and the PLUMED configuration needed to use it. The code has been tested with the latest version of Pytorch (1.4).

Submission history
[v1] 18 Feb 2020: original submission
[v2] 06 Apr 2020: updated doi and instructions

Click on the image below and get the code to add the badge to your website!