Erik Bernhardsson

Results 277 comments of Erik Bernhardsson

Yeah that probably makes sense. But you don't know that until after 8 minutes. And at that point, the 8 minutes are "sunk cost" anyway :)

Isn't page rank almost the same algorithm but just with a bit of regularization?

I'm pretty sure this is equivalent to Tikhonov regularization: https://en.wikipedia.org/wiki/Tikhonov_regularization And from a Bayesian point of view it's equivalent to putting normal priors on the stationary distribution. I could be...

i'm curious to see how much it changes the results... probably a little bit. let me run a side by side comparison and check

sorry, not sure what's going on :(

sorry – i don't have time to look into this. you are on your own :)

sorry i don't have any pointers (haven't looked at the code for like 2 years)

great if you can send a pull request if you get it working! thanks

is it binary compatible? i mean can you load an index generated in annoy / annoy-java?