The Gaussian approximation potential (GAP) is a machine learning approach to interpolating interatomic potential energy surfaces, based on Gaussian process regression. GAP is used in computational modeling of materials and molecules. GAP is also the name of the code used to perform these calculations.

On this website you can find current information and news about the code and methodology, as well as the GAP Developers & Users meetings organized by the GAP developers and its community.

The official repository for the GAP code is on Github:

For a brief introduction to the GAP methodology and an overview of the most relevant literature, check the What’s GAP page.

Latest news
Stay tuned for the GAP/(m)ACE D&U meeting 2023
Stay tuned for the GAP/(m)ACE D&U meeting 2023Miguel CaroMarch 31, 2023After the successful stint in 2022 in Finland, the GAP Developers and Users Meeting will be back in 2023 at Warwick University, UK with big news. In addition to all things GAP, the 2023 D&U meeting will also welcome developments in methodology, software and applications within the atomic cluster expansion (ACE) technique and its message-passing variant, MACE. Stay tuned for more news in this website! [...] Read more...