Neural networks predict the parity of Frobenius traces of elliptic curves from other traces with high accuracy, and their embeddings reveal learned mod-2 and mod-4 structure.
A toy model of universality: Reverse engineering how networks learn group operations
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.NT 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Learning Euler Factors of Elliptic Curves
Neural networks predict the parity of Frobenius traces of elliptic curves from other traces with high accuracy, and their embeddings reveal learned mod-2 and mod-4 structure.