Conditional Idempotent Generative Networks

TL;DR

Conditional Idempotent Generative Networks (CIGN) extend IGNs by enabling controlled, conditional data generation through new architectures and theoretical foundations, demonstrated on MNIST with promising results.

Contribution

Introduction of CIGNs that incorporate conditioning mechanisms into IGNs, with theoretical analysis and two architecture options for controlled generation.

Findings

Effective conditional generation demonstrated on MNIST

Two architectures (channel and filter conditioning) are viable

Foundation laid for future larger-scale experiments

Abstract

We propose Conditional Idempotent Generative Networks (CIGN), a novel approach that expands upon Idempotent Generative Networks (IGN) to enable conditional generation. While IGNs offer efficient single-pass generation, they lack the ability to control the content of the generated data. CIGNs address this limitation by incorporating conditioning mechanisms, allowing users to steer the generation process towards specific types of data. We establish the theoretical foundations for CIGNs, outlining their scope, loss function design, and evaluation metrics. We then present two potential architectures for implementing CIGNs: channel conditioning and filter conditioning. Finally, we discuss experimental results on the MNIST dataset, demonstrating the effectiveness of both approaches. Our findings pave the way for further exploration of CIGNs on larger datasets and with more powerful computing resources to determine the optimal implementation strategy.