The custodian engine surveys accumulated unclassified or generic claims across past optimization runs to detect coherent emerging patterns and mint new candidate rank factors for human-in-the-loop review.
Referenced by
