A simple neural model for unlabelled data

Code

• github.com/unverciftci/toUniform
  In papernone★ 6

Abstract

Neural networks are powerful tools for modelling big unstructured data. Advanced architectures and learning methods are used in applications. Nevertheless simpler models which use less data could be helpful. We propose a very simple learning algorithm in order to model the data distribution. Experiments show how effective and data-efficient our method is. Code is available at https://github.com/unverciftci/toUniform.

Tasks

Reproductions