Review of Disentanglement Approaches for Medical Applications -- Towards Solving the Gordian Knot of Generative Models in Healthcare

TL;DR

This paper reviews how disentanglement in generative models can improve interpretability, data privacy, and discovery of medical data relationships, emphasizing recent advances and applications in healthcare.

Contribution

It provides a comprehensive overview of generative models, disentanglement notions, evaluation metrics, and discusses their significance in medical applications.

Findings

Disentanglement enhances interpretability of medical models.

Generative models can synthesize privacy-preserving medical data.

Disentanglement aids in discovering novel medical data relationships.

Abstract

Deep neural networks are commonly used for medical purposes such as image generation, segmentation, or classification. Besides this, they are often criticized as black boxes as their decision process is often not human interpretable. Encouraging the latent representation of a generative model to be disentangled offers new perspectives of control and interpretability. Understanding the data generation process could help to create artificial medical data sets without violating patient privacy, synthesizing different data modalities, or discovering data generating characteristics. These characteristics might unravel novel relationships that can be related to genetic traits or patient outcomes. In this paper, we give a comprehensive overview of popular generative models, like Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Flow-based Models. Furthermore, we summarize the different notions of disentanglement, review approaches to disentangle latent space representations and metrics to evaluate the degree of disentanglement. After introducing the theoretical frameworks, we give an overview of recent medical applications and discuss the impact and importance of disentanglement approaches for medical applications.