arXiv · 2108.02221
Deep multi-task mining Calabi-Yau four-folds
Abstract
We continue earlier efforts in computing the dimensions of tangent space cohomologies of Calabi-Yau manifolds using deep learning. In this paper, we consider the dataset of all Calabi-Yau four-folds constructed as complete intersections in products of projective spaces. Employing neural networks inspired by state-of-the-art computer vision architectures, we improve earlier benchmarks and demonstrate that all four non-trivial Hodge numbers can be learned at the same time using a multi-task architecture. With 30% (80%) training ratio, we reach an accuracy of 100% for $h^{(1,1)}$ and 97% for $h^{(2,1)}$ (100% for both), 81% (96%) for $h^{(3,1)}$, and 49% (83%) for $h^{(2,2)}$. Assuming that the Euler number is known, as it is easy to compute, and taking into account the linear constraint arising from index computations, we get 100% total accuracy.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Harold Erbin, Riccardo Finotello, Robin Schneider, Mohamed Tamaazousti. 2021-08-04. Deep multi-task mining Calabi-Yau four-folds. https://doi.org/10.1088/2632-2153%2Fac37f7
Cite the original work for its findings. Save a collection to share your selection of sources.