Ryan Curtin

Ah, can I ask you to re-run these simulations? Could you run each one 10 times with different random seeds, and report the mean accuracy, variance of the accuracy (or...

Hmm, there's no variance in the result accuracy. Are you sure this is running correctly with different random seeds?

Right, I saw that, but no matter what there should definitely be some amount of variance in the accuracy---I don't expect that LMNN would converge to the same result with...

Right, looks good, could you also give results for covertype-5k and letters_recognition? For the accuracy results it looks like as long as the rank is "high enough" (which for iris...

Sure, it's ok if it takes some time. Are you sure the maximum number of iterations is being properly passed and reached? I can't see any reason why the results...

Hmm, I guess the different low-rank transformations could cause significantly different structure in the data which could then cause significantly different nearest neighbor search runtimes. I would not expect that...

Hi Arjun, There isn't currently any planned support for Ruby. If you are interested in writing a Ruby binding generator, it would be great and @zoq has given a link...

We need to wait for @zoq's input also, but my opinion is that any CLI program needs the following functionality: * specify a neural network architecture arbitrarily (possibly by loading...

> YAML for the config file sounds like a good idea to me, however parsing isn't as simple as e.g. csv and I don't like to include another dependency (perhaps...

This all sounds good to me. Probably we will have to flesh out the file configuration format somewhat, but if we can keep it of the form ``` ``` I...