Brendan Tracey
Brendan Tracey
@d4l3k I don't think the approach you have there is a good one. What you're doing is modifying the call such that `f(out of bounds)` returns `f(at bounds)`. From the...
Are you sure you have your signs correct? I can't tell for sure since I don't see where they are combined with your standard objective function. However, your barrier function...
If it's true that `Local` is returning minima that have a value of `inf` (i.e. outside the barrier) , then this is definitely a bug, and if you could make...
It's not clear to me yet, but I am adding the label because I think it may require a change.
I posted some thoughts, which probably should have gone on this issue but are here instead https://github.com/gonum/gonum/issues/925#issuecomment-482805916
I don't think that algorithm is computing the same object. This is computing a one-dimensional output, while the function for Uniform computes the multi-dimensional CDF ( https://en.wikipedia.org/wiki/Cumulative_distribution_function#Multivariate_case ). It could...
I'm not sure how the actual Gonum code is off the top of my head, but I've run into this a lot with probability distributions, so something to consider On...
The long-term plan is to provide support for complex matrices. I agree that the support is sparse right now (no pun intended). We would like pull requests to help expand...
Currently a mostly complete sketch of how it could look (but needs better comments and tests). I think a good idea would be to have a `running.Stats` that computes the...
I'm sorry, the last thing I pushed up was not what I meant to push. I pushed up what I intended now. Doing a refactor which will make it simpler.