blog:2026-09-24-ii
2026-09-24 (II)
last time, we ended at: if the input's intersections are within certain range, and all the possible output's intersections are present, then we have associative product. why is that? if we can show that all the newly created conicalization output intersections are impossible, then we should be able to do it.
before we do conicalization, the interior of the base Lagrangian has primitive value pretty large, positive or negative.
blog/2026-09-24-ii.txt · Last modified: 2026/09/25 08:13 by pzhou