Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

191 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Probabilistic Machine Learning + Deep Learning Project at the University of Trieste, 2023-2024

This project implements a process to perform anomaly correction of categorical data using two methods. One method uses the gradients of a classifier with respect to the inputs to transform a data input into a non-anomalous point. The other method leverages the distribution of the data learned by a diffusion model to correct data into a "healthy form". We also combined the two approaches into a pipeline to enhance the process. The first method produces a mask of the categories that could be modified using the distribution of the data learned by the diffusion model and then produces multiple versions of possible corrected instances.

CategoricalAnomalyCorrection

About

Repository with the joint project of the Probabilistic Machine Learning + Deep Learning courses at UniTS (2023-2024).

Resources

Stars

1 star

Watchers

1 watching

Forks

Releases

Packages

Used by

Contributors

Languages