TrIP2: Enhancing Transformer-Based Interatomic Potentials with Expanded Molecular Diversity for Accurate Energy and Force Predictions

The prohibitive computational cost of high-level ab initio quantum mechanical calculations has driven the recent development of machine learning interatomic potentials (MLIPs) that can efficiently approximate energies and forces with much lower computational expense. Traditional methods, such as density functional theory (DFT) or coupled-cluster theory, while highly accurate, become impractical for large systems or for applications requiring repeated evaluations, such as molecular dynamics simulations or large-scale materials screening. In response, MLIPs have emerged as a solution, offering orders-of-magnitude faster predictions while retaining quantum-level accuracy. However, many existing MLIPs suffer from limited generalizability, particularly when confronted with chemically diverse systems or novel configurations beyond their training sets. Here we introduce TrIP2, an advanced version of the Transformer Interatomic Potential (TrIP) trained on the expanded ANI-2x dataset, including more diverse molecular configurations with sulfur, fluorine, and chlorine. It leverages the equivariant SE(3)-Transformer architecture, incorporating physical biases and continuous atomic representations. Benchmarking on COMP6 energy and force calculations, structure minimization tasks, and QC torsion energy profiles, as well as applications to molecules with unexpected conformational energy minima, demonstrates TrIP2's high accuracy and transferability, outperforming or matching ANI-2x, the original TrIP, and AIMNet2. Notably, TrIP2 achieves state-of-the-art force prediction performance on the COMP6v2 benchmark, with a root mean square error (RMSE) of 1.07 kcal/(mol*Å). Without requiring any architectural modifications, TrIP2 successfully capitalizes on additional training data to deliver enhanced generalizability and precision, establishing itself as a robust and scalable framework capable of accommodating future expansions with minimal retraining. These attributes position TrIP2 as a highly promising tool for the development of transferable interatomic potentials.