You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is interesting. It appears to be a numerical stability issue caused by setting the observation noise too low. Currently, if you don't specify the logNoise then it defaults to -1e8. We've never had problems with this before, but now the variance (exp(2*logNoise)) seems to be rounded to 0.0 and this is causing issues with Optim. For now, I've changed the default noise on the master branch to -5.0 and we'll look into this in more detail.
For the time being, if you run
gp = GP(X,Y,mZero,kern,-5.0)
everything should work. Let me know if you're still having problems.
A domain error is triggered when running
I'm using
I've struggled to find a minimal example to trigger the error, the best I could do is for the values of X and Y given at the end.
It seems to me that the error comes from line 107 in GP.jl
gp.mLL = -dot((gp.y - μ),gp.alpha)/2.0 - logdet(gp.cK)/2.0 - gp.nobsv*log(2π)/2.0
when returns NaN when gp.cK has a negative determinant.
Running this code on my laptop where
generates a different error when running
optimize!(gp; method=Optim.ConjugateGradient())
I'm using the data
The text was updated successfully, but these errors were encountered: