Deep learning complete intersection Calabi-Yau manifolds
This work addresses the problem of handling algebraic topological data with machine learning for researchers in mathematical physics and string theory, but it is incremental as it builds on existing methods.
The paper reviews deep learning techniques for predicting Hodge numbers of complete intersection Calabi-Yau manifolds, achieving state-of-the-art accuracy and presenting new results on extrapolating predictions between low and high Hodge numbers.
We review advancements in deep learning techniques for complete intersection Calabi-Yau (CICY) 3- and 4-folds, with the aim of understanding better how to handle algebraic topological data with machine learning. We first discuss methodological aspects and data analysis, before describing neural networks architectures. Then, we describe the state-of-the art accuracy in predicting Hodge numbers. We include new results on extrapolating predictions from low to high Hodge numbers, and conversely.