The Newton-X platform for surface hopping and nuclear ensembles

Mario Barbatti, his group and collaborators published an update on Newton-X – a popular open-source platform for surface hopping and nuclear ensembles. An update include extension of the Newton-X platform to supervised (with our MLatom platform) and unsupervised learning (with ulamdyn). The paper is also open access and appeared in the Journal of Chemical Theory and Computation.

Our collaborative work summarizes capabilities and theory behind implementations of Newton-X, its design (it now branched into four stand-alone packages Initcond for nuclear ensembles, Newton-X classical and new series (upgraded faster version) for dynamics, and ulamdyn for data analysis), and describes many important updates. Here is the list of updates from Mario’s blog:

• zero-point-energy leakage correction;
• dynamics on complex-valued potential energy surfaces;
• dynamics induced by incoherent light;
• dynamics based on machine-learning potentials;
• exciton dynamics of multiple chromophores;
• supervised and unsupervised machine learning techniques.

We have been very happy to work with this great platform and we continue to contribute to it, namely, by implementing supervised techniques via interface to our MLatom package enabling simulating absorption spectra via machine learning-nuclear ensemble approach (see [1], [2], [3]) and nonadiabatic dynamics (see [1], [2]).

• Mario Barbatti*, Mattia Bondanza, Rachel Crespo-Otero, Baptiste Demoulin, Pavlo O. Dral, Giovanni Granucci, Fábris Kossoski, Hans Lischka, Benedetta Mennucci, Saikat Mukherjee, Marek Pederzoli, Maurizio Persico, Max Pinheiro Jr, Jiri Pittner, Felix Plasser, Eduarda Sangiogo Gil, Lijljana Stojanovic. The Newton-X platform: new software developments for surface hopping and nuclear ensembles. J. Chem. Theory Comput. 2022, ASAP. DOI: 10.1021/acs.jctc.2c00804.