Machine learning models achieve higher accuracy on prime patterns in Ulam spiral images from around 500 million than below 25 million, taken as evidence of greater learnable order at larger scales.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Machine Learnability as a Measure of Order in Aperiodic Sequences
Machine learning models achieve higher accuracy on prime patterns in Ulam spiral images from around 500 million than below 25 million, taken as evidence of greater learnable order at larger scales.