Of particular interest is the inter-atomic potential energy surface (PES).

Of particular interest is the inter-atomic potential energy surface (PES).

The paper proposes a new learnable model DeepPot-SE for inter-atomic potential energy surfaces (PES) based on deep neural networks. The authors start by introducing a number of requirements that a PES model should fulfil. Compared to other proposed model, this proposed model is the first to fulfil all these requirements, including differentiability and preserving natural symmetries. In the empirical evaluation, the performance of the proposed model is comparable to or better than state-of-the-art models as measured in MAE for energy and force predictions. Generally the paper the paper is well written and easy to follow.

Machine learning models are changing the paradigm of molecular modeling, which is a fundamental tool for material science, chemistry, and computational biology. Of particular interest is the inter-atomic potential energy surface (PES). Here we develop Deep Potential - Smooth Edition (DeepPot-SE), an end-to-end machine learning-based PES model, which is able to efficiently represent the PES for a wide variety of systems with the accuracy of ab initio quantum mechanics models. By construction, DeepPot-SE is extensive and continuously differentiable, scales linearly with system size, and preserves all the natural symmetries of the system. Further, we show that DeepPot-SE describes finite and extended systems including organic molecules, metals, semiconductors, and insulators with high fidelity.

Machine learning models are changing the paradigm of molecular modeling, which is a fundamental tool for material science, chemistry, and computational biology. Of particular interest is the inter-atomic potential energy surface (PES). Here we develop Deep Potential - Smooth Edition (DeepPot-SE), an end-to-end machine learning-based PES model, which is able to efficiently represent the PES for a wide variety of systems with the accuracy of ab initio quantum mechanics models. By construction, DeepPot-SE is extensive and continuously differentiable, scales linearly with system size, and preserves all the natural symmetries of the system. Further, we show that DeepPot-SE describes finite and extended systems including organic molecules, metals, semiconductors, and insulators with high fidelity.