Hello, Thanks for code & paper. I read the paper with interest. but i have one question about algorithm 1. i think row 6 localization map M^B extracted by classifier B. So I understood theta_A -> theta_B as S^~ as below, could you tell me why it came out with row6?