Skip to content

Compares neural network ensembles and Monte Carlo dropout in their ability to mimic the behavior of a Gaussian process.

License

Notifications You must be signed in to change notification settings

sisl/nns-as-gps

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

nns-as-gps

Comparison of neural network ensembles and Monte Carlo dropout in their ability to mimic the behavior of a Gaussian process.

The ensemble methods implemented are:

  1. A vanilla ensemble of neural networks, each predicting a single value, and trained using MSE loss.
  2. An ensemble of neural networks that each predict a mean and variance, trained with NLL loss, as advised by this paper.
  3. Monte-Carlo dropout, as described in this paper

Comparison of Methods

See also:

  1. Another implementation of ensembles with predictive uncertainty here

Author : Kunal Menda

About

Compares neural network ensembles and Monte Carlo dropout in their ability to mimic the behavior of a Gaussian process.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages