A Two-Step Disentanglement Method

TL;DR

This paper introduces a simple two-step adversarial training method to disentangle label-related factors from other data features, improving interpretability and data reconstruction across visual and financial datasets.

Contribution

The proposed method is simpler than previous approaches and effectively separates label-correlated factors from other data components using a two-step process.

Findings

Effective on visual datasets

Applicable to financial data

Simpler than prior methods

Abstract

We address the problem of disentanglement of factors that generate a given data into those that are correlated with the labeling and those that are not. Our solution is simpler than previous solutions and employs adversarial training. First, the part of the data that is correlated with the labels is extracted by training a classifier. Then, the other part is extracted such that it enables the reconstruction of the original data but does not contain label information. The utility of the new method is demonstrated on visual datasets as well as on financial data. Our code is available at https://github.com/naamahadad/A-Two-Step-Disentanglement-Method