{"language":[{"iso":"eng"}],"date_updated":"2024-01-05T12:57:31Z","title":"Towards active learning: A stopping criterion for the sequential  sampling of grain boundary degrees of freedom","publication":"arXiv:2302.01603","abstract":[{"lang":"eng","text":"Many materials processes and properties depend on the anisotropy of the\r\nenergy of grain boundaries, i.e.~on the fact that this energy is a function of\r\nthe five geometric degrees of freedom (DOF) of the interface. To access this\r\nparameter space in an efficient way and to discover energy cusps in unexplored\r\nregions, a method was recently established, which combines atomistic\r\nsimulations with statistical methods 10.1002/adts.202100615. This sequential\r\nsampling technique is now extended in the spirit of an active learning\r\nalgorithm by adding a criterion to decide when the sampling has advanced enough\r\nto stop. In this instance, two parameters to analyse the sampling results on\r\nthe fly are introduced: the number of cusps, which correspond to the most\r\ninteresting and important regions of the energy landscape, and the maximum\r\nchange of energy between two sequential iterations. Monitoring these two\r\nquantities provides valuable insight into how the subspaces are energetically\r\nstructured. The combination of both parameters provides the necessary\r\ninformation to evaluate the sampling of the 2D subspaces of grain boundary\r\nplane inclinations of even non-periodic, low angle grain boundaries. With a\r\nreasonable number of data points in the initial design, only a few\r\nappropriately chosen sequential iterations already improve the accuracy of the\r\nsampling substantially and unknown cusps can be found within a few additional\r\nsequential steps."}],"date_created":"2024-01-04T08:17:01Z","external_id":{"arxiv":["2302.01603"]},"_id":"50147","status":"public","year":"2023","author":[{"first_name":"Timo","full_name":"Schmalofski, Timo","last_name":"Schmalofski"},{"first_name":"Martin","last_name":"Kroll","full_name":"Kroll, Martin"},{"last_name":"Dette","full_name":"Dette, Holger","first_name":"Holger"},{"first_name":"Rebecca","last_name":"Janisch","full_name":"Janisch, Rebecca"}],"user_id":"67287","type":"preprint","project":[{"_id":"52","name":"PC2: Computing Resources Provided by the Paderborn Center for Parallel Computing"}],"citation":{"ieee":"T. Schmalofski, M. Kroll, H. Dette, and R. Janisch, “Towards active learning: A stopping criterion for the sequential  sampling of grain boundary degrees of freedom,” <i>arXiv:2302.01603</i>. 2023.","apa":"Schmalofski, T., Kroll, M., Dette, H., &#38; Janisch, R. (2023). Towards active learning: A stopping criterion for the sequential  sampling of grain boundary degrees of freedom. In <i>arXiv:2302.01603</i>.","bibtex":"@article{Schmalofski_Kroll_Dette_Janisch_2023, title={Towards active learning: A stopping criterion for the sequential  sampling of grain boundary degrees of freedom}, journal={arXiv:2302.01603}, author={Schmalofski, Timo and Kroll, Martin and Dette, Holger and Janisch, Rebecca}, year={2023} }","ama":"Schmalofski T, Kroll M, Dette H, Janisch R. Towards active learning: A stopping criterion for the sequential  sampling of grain boundary degrees of freedom. <i>arXiv:230201603</i>. Published online 2023.","chicago":"Schmalofski, Timo, Martin Kroll, Holger Dette, and Rebecca Janisch. “Towards Active Learning: A Stopping Criterion for the Sequential  Sampling of Grain Boundary Degrees of Freedom.” <i>ArXiv:2302.01603</i>, 2023.","mla":"Schmalofski, Timo, et al. “Towards Active Learning: A Stopping Criterion for the Sequential  Sampling of Grain Boundary Degrees of Freedom.” <i>ArXiv:2302.01603</i>, 2023.","short":"T. Schmalofski, M. Kroll, H. Dette, R. Janisch, ArXiv:2302.01603 (2023)."}}