Deep Coarse-grained Potentials via Relative Entropy Minimization

TL;DR

This paper demonstrates that relative entropy minimization improves the accuracy and data efficiency of neural network-based coarse-grained potentials, enabling better free energy surfaces and larger simulation time steps.

Contribution

It introduces the application of relative entropy minimization to neural network coarse-grained potentials, showing advantages over force matching in accuracy and efficiency.

Findings

RE training yields more accurate free energy surfaces.

RE is more data efficient than force matching.

RE corrects time integration errors, allowing larger time steps.

Abstract

Neural network (NN) potentials are a natural choice for coarse-grained (CG) models. Their many-body capacity allows highly accurate approximations of the potential of mean force, promising CG simulations at unprecedented accuracy. CG NN potentials trained bottom-up via force matching (FM), however, suffer from finite data effects: They rely on prior potentials for physically sound predictions outside the training data domain and the corresponding free energy surface is sensitive to errors in transition regions. The standard alternative to FM for classical potentials is relative entropy (RE) minimization, which has not yet been applied to NN potentials. In this work, we demonstrate for benchmark problems of liquid water and alanine dipeptide that RE training is more data efficient due to accessing the CG distribution during training, resulting in improved free energy surfaces and reduced sensitivity to prior potentials. In addition, RE learns to correct time integration errors, allowing larger time steps in CG molecular dynamics simulation while maintaining accuracy. Thus, our findings support the use of training objectives beyond FM as a promising direction for improving CG NN potential accuracy and reliability.