# Composition and decomposition of GANs

## TL;DR

This paper introduces a compositional framework for training GANs on data built from multiple components, enhancing modularity, interpretability, and extensibility of generative models, with theoretical identifiability conditions and empirical evaluations.

## Contribution

It presents a novel compositional training framework for GANs, enabling incremental construction and interpretability of complex generative models from simpler components.

## Key findings

- Framework improves modularity and interpretability of GANs
- Feasibility demonstrated on image and text datasets
- Provides conditions for model identifiability

## Abstract

In this work, we propose a composition/decomposition framework for adversarially training generative models on composed data - data where each sample can be thought of as being constructed from a fixed number of components. In our framework, samples are generated by sampling components from component generators and feeding these components to a composition function which combines them into a "composed sample". This compositional training approach improves the modularity, extensibility and interpretability of Generative Adversarial Networks (GANs) - providing a principled way to incrementally construct complex models out of simpler component models, and allowing for explicit "division of responsibility" between these components. Using this framework, we define a family of learning tasks and evaluate their feasibility on two datasets in two different data modalities (image and text). Lastly, we derive sufficient conditions such that these compositional generative models are identifiable. Our work provides a principled approach to building on pre-trained generative models or for exploiting the compositional nature of data distributions to train extensible and interpretable models.