Established in 2009, the Knowledgebase of Interatomic Models (KIM) is a cyberinfrastructure funded by the NSF aimed at improving the reliability of molecular simulations. It has the following features:
Curated repository of interatomic models (IMs) (potentials and force fields) with comprehensive provenance and version control.
Application Programming Interface (API) standards connecting molecular simulation codes (“simulators”) with IMs.
Standardized testing framework for archived IMs including their predictions for material properties and checks on their coding integrity.
Source and binary distribution framework for easy installation and use of the KIM API and KIM IMs with conforming simulators.
Rigorous transferability and uncertainty estimation for KIM IMs based on machine learning approaches to select IMs for target application and providing error bounds on their predictions (under development).
If you are new to KIM, start by reading the Getting Started Guide, which provides a brief overview of the main capabilities of the KIM framework and how to use them. You can also view an "Overview of the KIM Project," an online talk by Prof. Ellad B. Tadmor.