News for hackers. Stuff that matters.
Re:Minimum bipartite matching via Riemann optimization by amarcheschi on Wednesday January 08, 2025 @06:44AM (#42633200)
I had to deal with minimum bipartite matching when implementing FedMA (a federated learning algorithm). I was lucky it could be solved with a scikit learn method because i wouldn't have been be able to write a solution in a short time, let alone an efficient one

When my supervisor read about fedma using matching he went "oh god, i have terrible memories about it". he was right

Re:Minimum bipartite matching via Riemann optimization by marco_z on Wednesday January 08, 2025 @09:07AM (#42634351)
Yes, the problem can now be solved efficiently with a scikit one-liner, but the aim of this post was more speculative rather than beating the benchmark :)
Re:Minimum bipartite matching via Riemann optimization by amarcheschi on Wednesday January 08, 2025 @01:30PM (#42637027)
Oh yeah that's sure, i wasn't trying to diminish the content, the opposite. as in, it was so complex and out of what i expected that i was lucky to have a library that could solve it for me