Applications of machine learning in classical ciphers?

Question

Machine learning is definitely applicable in analyzing simple shift ciphers like Caesar and affine ciphers, as well as substitution ciphers like Vigenère, but is it possible for machine learning to solve problems involving more complex ciphers such as Hill ciphers, where characters are mixed up and ruin frequency analysis?

Answer 1

A cryptanalyst finds the right key for decryption, while a machine learning model finds a suitable solution in a large space of possible solutions. With enough data, the machine learning model could find a good solution. How much data is needed to reliably use machine learning for breaking complex ciphers such that it's less work than brute-force search could be a challenge.