Ryan Curtin

I'd like to keep this one open---it is really nice code, it just needs to be reviewed and merged. I hope to have the time sometime...

Right, I think this is close to correct. Here's a derivation, supposing x_i is the query point and x_a is the k-nearest neighbor (although the notation doesn't actually matter): ```...

Right, this is a bound for each individual query point, so I think the max() is not needed here.

I'll handle the tree-related part of this one, unless you are already working on it. I'm thinking about single-tree search first, and basically when we do k-nearest-neighbor search for a...

Yes, I have been thinking about it. I think that I can implement this one so you can focus on the other optimizations and BoostMetric. I'll open a PR when...

Right, I think that is not a bad idea. When you have a chance, be sure to go through and apply however many optimizations that you can that we've already...

An update on this one: I worked on it yesterday and today, and I have a branch rcurtin/lmnn-tree-rules that you can take a look at. It does not incorporate the...

Ah, ok, we can re-add that, it's no problem. I didn't realize that it was going to be used somewhere. :) And it is really not that much work for...

**Commented by speet3 on 18 Apr 41592779 02:53 UTC** (In #94) It appears I do not have the correct permission to modify the description of the ticket, so I'll just...

**Commented by speet3 on 3 Dec 41674892 09:35 UTC** Just as an FYI, this is somewhat of a two-part ticket. The build server can be made to run whatever profiling...