Mixture Density Networks: Probabilistic Regression for Uncertainty Estimation

 By Medium - 2021-03-20

Description

Uncertainty is all around us. It is present in every decision we make, every action we take. And this is especially true in business decisions where we plan for the future. But in spite of that, all…

Summary

  • Uncertainty is all around us.
  • Loss Function The network is trained end-to-end using standard backpropagation.
  • But if we use an MDN with a single component, it will approximate both the mean and the uncertainty in the function Image by Author Non-Linear Function Image by Author This is a non-linear function with a twist.
  • Image by Author Gaussian Mixture Image by Author Here we have two gaussian components which are mixed in a ratio, all of which is parameterized by x.
  • Image by Author The histogram also shows two small bumps on the top instead of a single smooth on of a normal gaussian.

 

Topics

  1. Machine_Learning (0.36)
  2. Backend (0.16)
  3. NLP (0.12)

Similar Articles

pytorch-widedeep: deep learning for tabular data

By Medium - 2021-02-22

This is the third of a series of posts introducing pytorch-widedeepa flexible package to combine tabular data with text and images (that could also be used for “standard” tabular data alone). The…

Automatic Social Share Images

By ryanfiller.com - 2020-12-11

I am a designer, developer, illustrator, and maker living and working in Memphis, Tennessee. This is my blog and portfolio.