Skip to content

scrambledpie/Conditional-Multi-Task-BayesOpt

master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 

Continuous Multi-Task Bayesian Optimization with Correlation

Toy code for reproducing the simpler experiments from the paper:

Continuous Multi-Task Bayesian Optimization with Correlation

The code is in R and is all self-contained in a single file. It requires R packages which are installed by running these command in the R console:

install.packages("FastGP")
install.packages("MASS")
install.packages("Rcpp")

The algortihm takes a function of two inputs and finds all the optima of one input conditional on the other input. In other words, optimize every vertical slice in the picture below, each slice is it's own Bayesian optiimzation problem, we just assume there is more than one, or even a continuum of problems.

@article{pearce2018continuous,
  title={Continuous multi-task Bayesian Optimisation with correlation},
  author={Pearce, Michael and Branke, Juergen},
  journal={European Journal of Operational Research},
  volume={270},
  number={3},
  pages={1074--1085},
  year={2018},
  publisher={Elsevier}
}

alt text

alt text

About

Toy code for reproducing experiments from the paper "Continuous Multi-Task Bayesian Optimization with Correlation"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages