search

gpstuff-dev / gpstuff

GPstuff - Gaussian process models for Bayesian analysis
http://research.cs.aalto.fi/pml/software/gpstuff/
GNU General Public License v3.0
169 stars 61 forks source link

Inverse-Distance Covariance Function #58

Closed faridf closed 1 year ago

faridf commented 1 year ago

Hello!

Thanks for making your code public. Is there any part of this code suppose to include the Inverse-Distance Kernel as it is suggested in this publication? Minimum Mode Saddle Point Searches Using Gaussian Process Regression with Inverse-Distance Covariance Function

https://pubs.acs.org/doi/10.1021/acs.jctc.9b01038

And if so, may I have access to it? Thanks!

avehtari commented 1 year ago

Hi, GPStuff is not actively developed and maintained any more, and that is why the inverse-distance kernel was not merged to this repo. If you are interested in using the minimum saddle point search for the energy surfaces, contact Hannes Jonsson who has led development of a more efficient code base which integrates also with the atomic simulation engine (ASE). If you want to replicate our experiments with the Matlab code, let me know, and I'll search the code for you.

faridf commented 1 year ago

Hi, thanks for your quick comment! I would be very glad to look into your Matlab codes if possible. Thanks!

avehtari commented 1 year ago

I found the Python code (for GPy), but not yet the Matlab code. I emailed Olli-Pekka.

faridf commented 1 year ago

May I actually have a copy of the GPy script as well. It could help me out since I am developing in python as well.

faridf commented 1 year ago

Also thanks for reaching out to Olli-Pekka! I appreciate it.

avehtari commented 1 year ago

The Matlab code is now at https://users.aalto.fi/~ave/atomic_GP-NEB-dimer_2020-07-09.zip The inverse-distance is in gpcf_sexpat.m

faridf commented 1 year ago

Thanks so much for sharing your code and I've been reading it and it's a great source for me.

I hope I'm not asking for much but but I could use the Python code that uses GPy as well. It would be very handy.

Thanks again!

avehtari commented 1 year ago

The Python code is now at https://users.aalto.fi/~ave/atomic_GPy-NEB-dimer_2020-07-09.zip The dimer part has not been tested and may contain errors.