Studying number theory with deep learning: a case study with the Mbius and squarefree indicator functions

Key Points

• The research aims to explore the use of deep learning techniques to compute the Möbius function and squarefree indicator function.
• Training small transformer models on number theory functions.
• Calculating the Möbius function and the squarefree indicator function.
• Using a mixture of additional models and feature scoring for theoretical insight.
• The models show nontrivial predictive power for the functions studied.
• The findings suggest deep learning models can effectively handle complex calculations in number theory.