#programming
This may end up being a dead end, but I've gotten some promising initial results on community detection using low-iteration Louvain against the follow graph weighted by interactions.
Time to do some release engineering.
We broke from our regular rhythm on this one to make it a slightly bigger update. It's going to be worth the wait though! It's got something for everyone.
If you're not a programmer, this is some fun trivia
If you are a programmer, feel free to correct anything I got wrong. I didn't get a CS degree so sometimes I botch big-O
RE:
After about 4 days of running non-stop, I've decided that O(nΒ³) time complexity Node Similarity calculation is probably a fool's errand. In hindsight I think 1.25Γ10Β²β° is too many computations.
A couple comments about what Iβve been learning about algo engineering, but first off: if youβre interested in helping us improve Discover and all of our algorithmic systems, we are hiring algo engineers [jobs.gem.com/bluesky/am9i...](
